Subscribe to Blind dev
<100 subscribers
Blind devGlider Fi: On-chain Portfolio and Rebalancing Strategies
Blind devAnalysis of Perp Dex Aggregator Liquid: Team, Concept, Coin, Code + Practice, Risks, and Advantages
Analysis of Perp Dex Aggregator Liquid: Team, Concept, Coin, Code + Practice, Risks, and AdvantagesAt first it was only a mobile app: I thought the project was useless for me. But it turned out there is a beta version of the web interface. So I decided to analyze the project.AuditTeamThere is no link to the team page or LinkedIn on the website. But I found out that the Liquid project was developed by a team of highly qualified specialists from New York with significant experience in quantitat...
Blind devEthos: a reputation service for X accounts and more. Analysis of the team, concept, coin, code, and …
Support Blind dev
Blind devZama Review: The Blockchain Bringing Full Privacy to Smart Contracts Across Any Network
Subscribe to Blind dev
<100 subscribers
Support Blind dev
Blind devZama Review: The Blockchain Bringing Full Privacy to Smart Contracts Across Any Network
Blind devGlider Fi: On-chain Portfolio and Rebalancing Strategies
Blind devAnalysis of Perp Dex Aggregator Liquid: Team, Concept, Coin, Code + Practice, Risks, and Advantages
Analysis of Perp Dex Aggregator Liquid: Team, Concept, Coin, Code + Practice, Risks, and AdvantagesAt first it was only a mobile app: I thought the project was useless for me. But it turned out there is a beta version of the web interface. So I decided to analyze the project.AuditTeamThere is no link to the team page or LinkedIn on the website. But I found out that the Liquid project was developed by a team of highly qualified specialists from New York with significant experience in quantitat...
Blind devEthos: a reputation service for X accounts and more. Analysis of the team, concept, coin, code, and …
Share Dialog
Share Dialog
Zama makes familiar blockchains private: now smart contracts can process encrypted data that is not even visible to nodes. This is a level of privacy without migrating to a new chain - Zama works on top of any L1 and L2. Demo applications are already available, including a confidential portfolio and other use cases.
Let's explore the project: how good is it.
There is information about the team in the "About" section and in more detail (albeit without LinkedIn links) in the Litepaper:
Rand Hindi, CEO of Zama. Work experience: Co-founder of Zama since 2020 to present; was a partner at Unit.vc; founded the startup Snips, which was acquired by Sonos; previously - member of the French Digital Council (advised the government on AI and Privacy issues); was a lecturer at Science Po University in Paris; advisor to several companies in biotech, AI and defense.
Education: BSc Computer Science and PhD Bioinformatics, University College London UCL.
Pascal Paillier, CTO of Zama. Work experience: co-founder of Zama since 2020; inventor of the Paillier additive homomorphic encryption scheme; before Zama, he headed the cryptographic innovations team at Gemalto; founded the consulting firm CryptoExperts; 2025 - IACR Fellow; led several ISO cryptography standards.
Education: PhD in Cryptography, Telecom Paris.
Nigel Smart, Chief Academic Officer of Zama. Work experience: one of the leading researchers and executives of the company; specific workplaces and dates are not specified.
Marc Joye, Chief Scientist of Zama. Work experience: one of the leading researchers and executives of the company; specific workplaces and dates are not specified.
Also found on the site LinkedIn:
10 thousand followers.
Last post a week ago. It received 46 reactions and 4 reposts. They wrote about the story of Zama and FHE.
As many as 170 participants associated with the company!
I think this is too many... But I may not know the specifics of developing such projects.
I selected a small list of key employees for analysis.
8,791 followers.
Shares other content in posts, but also about Zama.
Work experience
Co-founder & CEO at Zama since November 2019
General Partner at Unit Ventures since January 2020
Advisor at CrunchDAO since February 2022
Advisor at Mindstate Design Labs since June 2021
Advisor at Chipiron since February 2021
Advisor at NAVAL GROUP from September 2018 to December 2021
Lecturer at Sciences Po from September 2017 to December 2021
Advisor at Sonos, Inc. from January 2020 to January 2021
Total 14.
Education
At UCL: PhD, Bioinformatics from 2006 to 2011
At UCL: Bsc, Computer Science from 2003 to 2006
THNK School of Leadership: Graduate Degree, Creative Leadership from 2012 to 2013
Singularity University: Graduate program, Exponential Technologies in 2010
Good experience.
4,920 followers
No publications.
Work experience
Chief Technology Officer at Zama since November 2019
At CryptoExperts: Chief Executive Officer from February 2009 to December 2019; Senior Cryptography Expert from February 2009.
Non Executive Advisor at Nitromia since May 2019
Non Executive Advisor at CircaGene since April 2019
Non Executive Advisor at Olvid since March 2019
Co-editor of ISO/IEC 27551 "Requirements for attribute-based unlinkable entity authentication" at ISO since April 2017
Editor of ISO/IEC 18033-6 "Homomorphic Encryption" at ISO from October 2015 to October 2019
Total 11.
Education:
Telecom Paris: PhD, Computer Science, Cryptography from 1995 to 1999
Extensive experience. Great that he has worked in leadership positions and engaged in cryptography + FHE - this is useful for the project.
10,020 followers.
Last post 1 day ago with thanks for 10K followers with 54 reactions and 7 comments.
Previous - 6 days ago with 6 reactions. Post about tGBP integration, "Wirex."
Work experience:
Director of Business Development at Zama since February 2024
Membre at Crypto Valley Association since March 2024
Mentor at Outlier Ventures since May 2024
At Ledger: Inside Sales - Ledger Enterprise from January 2020 to December 2020; Sales Director - EMEA - Ledger Enterprise from January 2021 to December 2021; Sales Director - UKI & Nordics - Ledger Enterprise from January 2022 to January 2024
Sales Development Representative at Snips from 2019 to 2020
Business Developer at Médoucine from 2017 to 2019
Commercial Development / Pre-sales at Worldline in 2016
Education
Frankfurt School of Finance & Management: DeFi Talents, Decentralized Finance
Conservatoire National des Arts et Métiers: Fundamentals of Blockchain Technology, Finance, General
IÉSEG School of Management: MSc in Management - Grande école, International Business, Economics & Strategy
Good work experience and profile. Experience in Ledger Enterprise directly supports the current BizDev role.
4,023 followers.
Last post 1 month ago about Zama with 32 reactions and 3 reposts.
Work Experience
Ecosystem Lead at Zama since January 2023
At Theodo: Product Manager from March 2017 to March 2020; Head of blockchain BU and nocode BU from March 2020 to March 2022
Co-founder at Bannann from September 2015 to June 2018
Equity Derivative Sales at BNP Paribas from January 2014 to August 2015
Investor Relations Assistant at Anaxago from February 2013 to August 2013
Education
ESCP Business School: Master in Management Grande Ecole from 2011 to 2015
Queen’s University: 6-Months Exchange Program, Business / Commerce in 2014
Lycée Kléber: Classe préparatoire aux grandes écoles - Economique et commerciale option Scientifique (ECS) from 2009 to 2011
He has experience in product management and building blockchain ecosystems - this is useful for the current role of Ecosystem Lead.
2,272 followers.
Last post 2 months ago with 61 reactions. About Zama and privacy blockchains.
Work Experience:
VP - Corporate Development at Zama since October 2022
Strategy Management Consultant from April 2022 to October 2022
At Digital14: VP - PMO and Corporate Development from August 2015 to April 2017; SVP - Strategy, Performance Management and Corporate Development from April 2017 to October 2019; Chief Financial Officer and Head of Strategy from October 2019 to December 2021
Strategy Management Consultant from August 2014 to August 2015
CEO and Co-Founder at KooraBazar.com from January 2013 to July 2014
VP - Investment and Asset Management in ICT at Mubadala Development Company from January 2009 to January 2013
Engagement Manager at Booz & Company from November 2006 to December 2008
Line Manager at Axalto from July 2003 to July 2005
Total 11.
Education
INSEAD: MBA, Finance / Entrepreneurship / Strategy from August 2005 to July 2006
Stanford University: Master of Science (MS), Electrical and Electronics Engineering from September 1998 to June 2000
École Polytechnique: Diplome de l'Ecole Polytechnique, Maths / Electronics from September 1995 to June 1998
Lycee Prive Ste Genevieve: Maths from September 1993 to June 1995
Тоже хороший опыт у него.
3 973 отслеживающих.
Последняя публикация 1 неделю назад с 38 реакциями, 3 комментариями и 1 репостом. Про найм в Zama.
Опыт работы
В Zama: Marketing Director с октября 2021 по настоящее время; Vice President Marketing с апреля 2024 по настоящее время
Co-founder - Growth and acquisition в Siliconwood с ноября 2017 по сентябрь 2021
Ruby on Rails Developer (Freelance) с мая 2012 по декабрь 2020
Growth Hacker в Leetchi.com с августа 2014 по февраль 2017
Education
San Francisco State University: Digital Marketing and Entrepreneurship from December 2013 to July 2014
International School of Business, Dublin: Bachelor of Business Administration (B.B.A.) in 2012
ISTEC - Ecole Supérieure de Commerce et de Marketing: Master’s degree from 2010 to 2015
Ecole Suger: Baccalauréat scientifique in 2010
Longtime at Zama: held several marketing positions - a professional in this. And the project is popular, which proves it.
He has years of experience in marketing/growth (Growth Hacker, co-founder in growth, specialized training), so past roles confirm professionalism in marketing.
172 followers.
Last post 6 months ago with 10 reactions, 3 comments, and 3 reposts.
Work experience:
Senior Developer Relations at Zama from October 2024 to present
Technical co-founder at Lastic from September 2023 to present
Community Engagement Coordinator & Web Developer (Volunteer) at SubWork - coworking in nature from November 2022 to present
Co-Founder at Women Of Polkadot from October 2023 to December 2024
Founder at PoppySeed Dev from January 2023 to December 2024
- Total 15.
Education
Polkadot Blockchain Academy: Bootcamp, Blockchain Development from July 2023 to August 2023
DappCamp: Bootcamp, Blockchain Development from October 2022 to November 2022
University of Ljubljana, Faculty of Computer and Information Science: Master’s degree, Intermedia/Multimedia from October 2020 to February 2021
University of Ljubljana, Faculty of Electrical Engineering: Bachelor’s degree, Electrical and Electronics Engineering from October 2018 to July 2020
University of Ljubljana, Faculty of Mathematics and Physics: from 2016 to 2018
Good experience and education. Judging by the previous experience, may perform well in the current position at Zama.
425 followers.
Last post 1 week ago with 38 reactions, 3 comments, and 1 repost. About a community lead vacancy at Zama.
Previous post 3 weeks ago with 122 reactions, 6 comments, and 13 reposts. About Zama and OpenZeppelin partnership on confidential smart contracts.
Work experience
At Zama: Senior GPU Researcher from February 2022 to September 2022; Senior GPU Engineer from February 2022 to present
At Universidade Estadual de Campinas: Scientific initiation researcher from March 2010 to December 2011; Master of Science from February 2014 to June 2016; PhD candidate from February 2016 to March 2023
Visiting PhD Student at Aarhus University from September 2020 to September 2021
Business Analyst at Kanui from April 2013 to March 2014
Scientific initiation researcher at Nacional Laboratory of Biosciences LNBio/ABTLuS from March 2012 to July 2013
Education
Universidade Estadual de Campinas: Doctor of Philosophy (PhD), Computer Science from 2016 to 2022
Universidade Estadual de Campinas: Master of Science (MSc), Computer Science from 2014 to 2016
Universidade Estadual de Campinas: Bachelor's degree, Applied Mathematics from 2008 to 2013
Strong research and engineering background: applied mathematics → MSc/PhD in computer science, experience with GPU/parallel computing and academic research; has been at Zama for a long time (from researcher to senior engineer), indicating accumulated expertise and good relevance to the company's tasks.
99 followers.
Last post 3 weeks ago with 122 reactions, 6 comments, 13 reposts. About Zama and OpenZeppelin partnership on confidential smart contracts.
Previous post 1 month ago. Received 155 reactions, 7 comments and 26 reposts. About the announcement of Zama Protocol and the launch of the public testnet.
Work experience
Senior Solidity Engineer at Zama from April 2025 to present
Senior Lead Software Engineer, freelance from June 2021 to April 2025
Career break from 2020 to 2021
Co-Founder, Principal Software Engineer at GlobFX Technologies from 2012 to 2019
Co-Founder, Principal Software Engineer at GlobFX Technologies from 2000 to 2019
Co-Founder, Principal Software Engineer at GlobFX Technologies from 2004 to 2009
Java Software Developer at Dassault Systèmes from 1998 to 2000
Education
Grenoble INP Ensimag: Master’s degree, Computational and Applied Mathematics from 1993 to 1996
Lycée Henri IV: MPSI MP from 1992 to 1993
Good experience in development (engineering). Worked in this direction not only at Zama, but also earlier.
511 followers.
No publications.
Work Experience
Cryptographe at Zama from October 2020 to present
Chercheur en cybersécurité at WALLIX Group from December 2018 to October 2020
Doctorant sur le sujet "Evaluation de la confiance dans les architectures de sécurité" at Université Grenoble Alpes from October 2014 to August 2018
Enseignant (moniteur/ATER) at Université Grenoble Alpes from September 2015 to June 2018
Internship "Classification of semi-bodies and substitution tables for symmetric ciphers" at Laboratoire Jean Kuntzmann from February 2014 to July 2014
Industrial Project "Acceleration of algorithms for digital printing" at Caldera from January 2013 to July 2013
Education
Université Grenoble Alpes: Doctorate, Mathematics and Computer Science 2018
Grenoble INP Ensimag: Specialization in Security, Cryptography and Information Coding 2014
Grenoble INP Esisar: Engineering Degree, Computer Science and Networks 2014
Lycée Pierre d'Aragon: Scientific Baccalaureate with Mathematics specialization 2009
Professional: foundation in cryptography and security, strong skills in formal proofs and protocol design, focus on verifiability and practical applicability of solutions.
In the blog the latest post on July 30 - announcement of partnership:
Developer forum active:
Last message 16 minutes ago (August 14 at 16:44 MSK).
Active chat (wrote a minute ago). Questions are answered.
And correctly.
At first I thought they weren't getting through, but now I've reached the community section on the website, and there's a link to LinkedIn.
In announcements, the last message was on August 12 (normal):
In the creators' program chat, they also write actively:
In technical support as well:
In the developers' program chat as well:
Overall, Discord is super active!
X @zama_fhe:
215.9K subscribers.
1,699 posts. Last one on August 8 about community update (developer program winners). It got 106 replies, 115 reposts, 439 likes, and 33,427 views.
Previous one was on August 5 about meeting project founders to discuss HFE and Zama. It got 80 replies, 108 reposts, 438 likes, and 18,553 views.
X analysis by Tweetscout:
Score 771, level 3 (Credible).
Subscribed projects: @MilkRoadDaily, @flipsidecrypto, @OpenZeppelin, @protocollabs, @JSCCapital, @GSR_io, @ShytoshiKusama, @conduitxyz and others.
Venture capital: @tarunchitra, Robot Ventures, @jbrukh, CoinFund, @VinnyLingham, Newtown Partners and others.
Analysis of X by Moni:
Moni ScoreLevel: 4. Established
3551
0 per day.
136 mentions by smarties, 290 smarties.
According to Moni, the following projects are subscribed: @NEARProtocol, @Filecoin, @Shibtoken, @SecretNetwork, @OasisProtocol, @PhalaNetwork, @Try_Wink, @vita_dao and others.
VC: @robotventures, @GSR_io, @hack_vc, @Signum_Capital, @StakeCapital, @ej__rogers.
Good level of fame and activity for X.
64,107 subscribers, last post on August 2 with 8K views and > 200 reactions - that's cool!
3,93K subscribers. Last video two weeks ago with 221 views.
Last messages 1 month ago. Many reactions - > 200-300.
11 thousand subscribers. Last post 15 days ago with 49 reposts, 2 quotes, 177 likes and 2.4K views.
Previous 22 days ago with 27 reposts, 1 quote, 99 likes and 1.2K views.
Rating 5 out of 5: many team members, and the main ones (analyzed) are professionals. Social media is active (even Youtube with Reddit) and activity is good.
They answer questions in Discord and correctly.
There is documentation, and as part of it - Litepaper:
We considered the problem: blockchains are needed for finance, governance, and identification. But they are not confidential (all data is public). This is a problem. No large or medium-sized investors want everyone to see the amount of money and where it is located.
Zama is an interchain privacy layer on top of any L1/L2 that adds confidential smart contracts with programmable access rules to existing blockchains. Input data and transaction states are end-to-end encrypted, not even visible to nodes, while contracts remain composable with both private and public protocols.
Since it's not a new L1/L2, users don't need to migrate to another chain, and they can work with private dApps from anywhere.
Zama is based on fully homomorphic encryption (FHE), which allows computations to be performed directly on encrypted data without revealing it. The results remain publicly verifiable.
Zama's FHE technology is inherently post-quantum, already accelerated 100+ times, and supports familiar languages like Solidity and Python. The team calls it the next step after HTTPS - "HTTPZ", where privacy becomes the base for all on-chain applications.
For scalability and flexibility, the protocol combines FHE with MPC and ZK: keys are decentralized, inputs are efficiently verified, and performance scales with GPU and specialized hardware.
As a result, Zama creates a universal privacy layer that doesn't require changing the blockchain but is compatible with any L1 and L2.
The Zama protocol is based on years of research and development by Zama. The public testnet is already running, allowing everyone to deploy and test private dApps, and operators to set up coordination and processes. The first official mainnet is scheduled for Q4 2025 on the Ethereum blockchain: it will bring privacy to the Ethereum ecosystem, and only Zama-approved applications will be available at launch. By the end of the year, there will be a TGE with the launch of the $ZAMA token and connection of other EVM chains to the Zama protocol, after which dApp deployment will be possible without permission. In 2026, after completing the initial EVM-only phase, the protocol will be deployed in Solana, opening up confidential SVM applications.
It's great that they identified them, as it helps understand the demand scenarios for Zama.
Confidential smart contracts open up a new design space for on-chain applications, especially in finance, identification, and governance. In the web2 world, most services do not publicly disclose data, which means the lion's share of on-chain applications is yet to come thanks to Zama.
In finance, the protocol enables private payments: balances and transfer amounts are end-to-end encrypted, and compliance requirements can be embedded directly into token contracts. For tokenizing real-world assets, large institutions can use public blockchains like Ethereum and Solana while maintaining the confidentiality of operations and investor identities and performing KYC/AML in smart contracts without disclosing sensitive data. This is confirmed by a PoC in the JP Morgan - Kynexis report based on Zama technologies. In DeFi, fully encrypted swaps become possible with protection against front-running and hiding amounts and even assets, as well as private lending, on-chain credit scoring, and option pricing.
In token operations, closed auctions with sealed bids appear: participants publish encrypted bids, and only the winner is revealed at the end, improving pricing and protecting against bots from the mempool, especially in public token sales. Distributions can be conducted confidentially without disclosing amounts on the address, supporting vesting on encrypted tokens and private staking.
For identification and governance, composable on-chain identity is supported: a full DID + VC scheme lives on the blockchain in encrypted form and remains compatible with dApps, providing an abstraction of id and allowing smart contracts to verify claims for regulatory compliance privately and decentrally.
Voting becomes confidential: individual votes and stakes are hidden, only the result is revealed.
Other examples include on-chain corporations, where the capital structure, finances, board decisions, customer and employee data remain private, and smart contracts automate operational processes.
Prediction markets get encrypted forecasts with periodic disclosure, reducing biases and increasing accuracy.
Data markets for AI allow users to selectively share and sell data, train models on encrypted datasets, and receive a steady stream of income instead of a one-time sale.
This is just a part of what is already available today. As FHE scales through the Zama protocol, liquidity will grow and the transition of users and companies on-chain will become easier. Over time, it will be possible to run entire companies, cities, and even countries on the blockchain with financial and identification infrastructure, elections, currency, taxes, and registries of land, vehicles, and companies. Confidential blockchains provide not only programmable money but also programmable public infrastructure.
Most chains cannot natively perform FHE.
Zama solves this with two ideas: symbolic execution and threshold decryption.
Symbolic execution
A contract on the host chain invokes FHEVM. The chain itself does not compute FHE: it writes a pointer to the result and emits an event.
Coprocessors pick it up. They perform FHE computation and store the ciphertext.
Advantages:
no need to change the chain or install special hardware,
non-FHE transactions are not slowed down,
FHE operations can be parallelized.
Since the host chain stores pointers, not the ciphertexts themselves (data is stored by coprocessors), FHE calls are easily composable without waiting for previous ones to complete (waiting is only needed at the decryption moment).
Security:
From a security standpoint, all coprocessor actions are publicly verifiable: anyone can recalculate the ciphertexts and verify their correctness. At launch, several coprocessors with majority consensus are used, and in the future, it is planned to allow anyone to perform FHE operations with proof of correctness via ZK-FHE.
Threshold Decryption
To maintain on-chain composability, all ciphertexts are encrypted under the same public key, meaning the private key must be protected from unauthorized decryption. In the Zama protocol, the private key is split between participants and managed by a dedicated threshold MPC as a key management service (KMS). A user or contract can only decrypt a value if the source contract explicitly granted such permission on the host chain; a decryption request is sent to the Zama Gateway, which orchestrates the process and redirects it to KMS participants. All such requests are publicly visible, so anyone can verify their compliance with the access control logic defined by smart contracts.
Components
The ecosystem includes several basic components.
Host chains are L1/L2 protocols supported by Zama, on which developers deploy confidential dApps.
The FHEVM library provides them with tools to create private smart contracts, and the FHEVM Executor contract on the host chain accepts dApp calls, records FHE operations, and automatically emits events for coprocessors.
An access control list (ACL) is deployed on each host chain and tracks who is authorized to compute over an encrypted value and who can decrypt it; coprocessors relay events from the ACL to the Gateway, where ACLs from different host chains are aggregated into a single registry for authenticating KMS decryption requests.
The native $ZAMA token is used to pay fees, stake, and govern.
Gateway is a set of smart contracts that orchestrate the protocol: users send encrypted inputs for verification, request decryption, and bridge encrypted assets between host chains through it, paying a small fee in $ZAMA.
For maximum performance and efficiency, a dedicated rollup on Arbitrum is chosen: it serves only the Zama protocol and does not allow deployment of third-party contracts.
Coprocessors are nodes that verify encrypted inputs, perform FHE computations, store resulting ciphertexts, and relay ACL events to the Gateway; each node commits its result, after which the Gateway makes a decision by majority. All their tasks are publicly verifiable, and scaling is possible both vertically and horizontally.
KMS nodes run MPC protocols for key generation, CRS, and threshold decryption; no single party ever has access to the keys. KMS nodes are orchestrated via Gateway and operate within AWS Nitro Enclaves, which complicates key share leakage and provides software version attestation; ZK-MPC is planned to be added for verifiability without relying on hardware.
Operators are organizations that run protocol nodes, including coprocessors and KMS.
Performance
The Zama protocol was initially designed to be horizontally scalable and relies on the TFHE-rs library. Unlike the sequential EVM model, FHE operations are computed in parallel, and if a specific ciphertext is not involved in a sequential chain, coprocessors increase throughput by simply adding servers.
Since the start of protocol development, throughput has increased from 0.5 to 20+ transactions per second per host chain, which in total across all host chains theoretically provides hundreds of tps. This is sufficient for most EVM chains (Ethereum ~15 tps), but insufficient for retail payments (VISA ~25,000 tps) and for Solana (1000+ tps).
The gap is planned to be closed gradually through hardware: transitioning from CPU to GPU will provide ~50-100 tps per chain (for state-independent transactions), an open FPGA accelerator aims for 500-1000 tps per chain, and specialized ASICs - for 10,000+ tps per chain.
The key point is that FHE limits are now primarily defined by 'Moore's Law' - the better the hardware, the higher the protocol's throughput.
Security
The approach is defense-in-depth with a combination of multiple layers of protection. All FHE operations use 128-bit cryptographic strength and a p-fail probability of 2^-128; the scheme is post-quantum, meaning it is resistant to quantum attacks. All FHE computations are publicly verifiable: anyone can recalculate the results and detect dishonest FHE nodes; this is similar to the security of optimistic rollups, but applied to FHE. At the same time, the results are signed not by one, but by three independent FHE nodes from different operators, providing both 'optimism' and consensus at the output.
The MPC protocol involves 13 nodes with a 2/3 majority rule, and the protocol itself works and is correct even with up to 1/3 malicious participants - according to the team, this is the first production implementation of a working MPC.
MPC runs in AWS Nitro Enclaves, adding a layer of protection and preventing access to the share of the private FHE key from outside; software attestation allows the protocol to track updates. The combination of MPC and Nitro means that compromising keys would require collusion between AWS and several MPC nodes.
Genesis operators are public and reputable organizations with billions at stake in their core business (validators, infrastructure providers, etc.); being doxxed, they bear reputational and economic risks for any violation.
Slashing is managed through governance, allowing for proposing sanctions and addressing edge cases specifically.
The protocol is undergoing one of the largest audits in the industry by Trail of Bits and Zenith - totaling over 34 audit weeks.
Compliance
Compliance rules are defined in smart contracts.
The protocol does not decide who can decrypt what, but the application developers do.
Development plans
Achieve 10,000+ tps:
new FHE techniques with 10-20x acceleration,
FHE ASIC with 100-1000x acceleration,
migrate Gateway to ZK rollup with less than 100 ms delay.
Strengthen KMS:
ZK-MPC for trustless correctness verification,
more nodes in the MPC committee - dozens and hundreds.
Open participation to operators:
launch MPC inside HSM,
ZK-FHE, where correctness is proven, and coprocessors compete like in PoW.
Currently expensive, but progress is being made.
Full post-quantum stack:
Replace ZKPoK with lattice-based post-quantum, host-chain signatures are not post-quantum yet.
L1/L2 community migration.
The Zama protocol uses Delegated Proof-of-Stake.
It is run by 16 operators - initially 13 KMS nodes and 3 FHE coprocessors (the number will grow over time).
Operators are selected according to the following rules:
Genesis operators are selected based on reputation, DevOps experience, and offchain value (equity, revenue, market cap, etc.). This provides initial security through reputation.
A large operator risks losing customers if caught violating the Zama protocol.
Eventually, anyone will be able to become a KMS operator or coprocessor. You need to reliably demonstrate node operation in the testnet, and then stake at least 0.5% of the circulating supply of $ZAMA. Each epoch (initially 3 months), the protocol selects the top 13 KMS and top 3 coprocessors by stake for the next epoch.
All 16 active operators receive rewards in $ZAMA. The size depends on the role and stake share.
The team is currently looking for genesis operators.
If you want to apply - fill out the form.
Token holders without the necessary infrastructure can participate through Delegation.
They delegate their $ZAMA to operators from the whitelist and receive a portion of their rewards.
Each operator independently chooses the motivation for delegators - reduced commission or additional rewards not in $ZAMA.
Most operators adopt protocol updates. This includes software updates, commission changes, adding new host chains, etc.
Exception - emergency protocol pause and spammer blacklist, any operator can do this.
Unpausing or removing from the blacklist can only be done through regular social consensus.
Slashing is provided for abuse. This way the protocol quickly responds to critical cases and encourages honest behavior.
On-chain voting by $ZAMA holders is used for two things:
changing inflation,
slashing operators who break the rules.
There is an official dashboard:
Transaction fees (total) - 1,082 Sepolia ETH
529,046 transactions
108,822 users
6,979 contracts
Rating 4 out of 5: everything is described in detail. There is information about the reasons for the demand for Zama (finance, identification, and management cannot be public, as in the web2 world, not everyone can see the data. Accordingly, the success of applications on Zama is yet to come), and technical details.
The idea is also unique: a level for inter-blockchain privacy. But the downside is that the network is currently partially centralized and is unlikely to provide good scalability compared to existing blockchains without centralization by transitioning to ASICs.
The litepaper also contains information about the token
$ZAMA is the native token of the protocol. It is needed for fees, staking, and governance.
The burn-and-mint model is used - 100% of fees are burned, new tokens are minted for operator rewards.
Fee model
Deploying a confidential application on a supported network is free and permissionless. The Zama protocol does not charge a fee for FHE calculations.
Payment is charged for three actions:
ZKPoK verification. The user pays when adding encrypted inputs to a transaction. The price depends on the number of bits being verified.
Decryption of ciphertexts. The user pays for each decryption request. The price depends on the number of bits to be decrypted.
Bridging of ciphertexts. The user pays for transferring an encrypted value between networks. The price depends on the number of bridged bits.
The commission can be paid by the end user, frontend, or relayer.
Developers can make it so that users don't need to hold $ZAMA.
Fees are paid in $ZAMA but denominated in USD.
The oracle regularly updates the $ZAMA/USD price on the Gateway.
Gateway contracts recalculate how much $ZAMA is needed for each function.
This creates two effects:
fees are tied to usage, not speculation,
it's easier for users, developers, and relayers to plan expenses in USD.
Volume discount applies:
The more an address verifies/decrypts/bridges in 30 days, the lower the price per bit.
Discounts range from 10% to 99% depending on volume.
Starting price grid:
ZKPoK verification: $0.016 to $0.0002 per bit,
Bridging: $0.016 to $0.0002 per bit,
Decryption: $0.0016 to $0.00002 per bit.
Decryption is 10 times cheaper than ZKPoK and bridging. This is due to frequency - a 10:1 ratio in favor of decryptions is expected.
Example of confidential token transfer:
amounts and balances take up 64 bits,
usually 3 decryptions are needed per transaction - sender's balance, recipient's balance, and final amount (0 if transfer failed),
then the final cost, including discounts:
ZKPoK verification: 64 * ZPoK_bit_price = $0.01 to $1,
decryption of 2 balances + amount: 64 * 3 * decryption_bit_price = $0.003 to $0.30,
total: about $0.01 to $1.3 on average.
The model makes operations accessible to large users and profitable for operators.
A small occasional user will pay about $1 per transaction. A large user (e.g., wallet), on the other hand, will pay up to $0.01 per transaction.
Base prices are set so that 1 tps on a single host chain yields about $3.5m in fees per year.
If 10% of global blockchain transactions use Zama for privacy (about 300 tps), the protocol's annual revenue could reach $1 BILLION.
Staking rewards
Operators stake $ZAMA to participate in the protocol and receive rewards. They are minted by inflation - initially 10%.
The rate can be changed on-chain.
Distribution is by roles - sequencer, coprocessors, KMS. Then rewards are divided proportionally to stake within the group.
The operator decides how to share with delegators - expected commission of about 20%.
Expected parameters by groups:
Coprocessors - 3 operators - about 13.3% rewards per operator. Cost estimate: $20,000 per month for every 10 tps host chains.
KMS - 13 operators - about 4.6% rewards per operator. Cost estimate: $2,500 per month for 50 tps decryptions.
This scheme rewards contributions based on actual work done. It also motivates delegators to carefully choose who to delegate to.
Distribution
No details yet.
According to Cryptorank, raised $130 MLN with a valuation of $1 MLRD!:
Investors: Pantera Capital, Multicoin Capital, Anatoly Yakovenko, Protocol Labs, Blockchange Ventures, Stake Capital Group, Metaplanet, Vsquared Ventures, Juan Benet and Gavin Wood.
Rating 4 out of 5: token utility exists, but no initial distribution yet. Investments are excellent (130 MLN $ from well-known investors, e.g., Pantera Capital, Multicoin Capital and Anatoly Yakovenko).
There are GitHub repositories:
There are 64 of them, so we'll look at the main ones.
:
FHEV, a full-featured framework for integrating fully homomorphic encryption (FHE) with blockchain applications.
Last commit 44 minutes ago:
Total 3,245 commits:
August 14, 13, 12, 11, 8, 7, 6, 5, 4, 3, 1 and earlier.
:
TFHE-rs: Pure Rust implementation of TFHE scheme for logical and integer arithmetic on encrypted data.
Last commit 6 hours ago:
Total 3,192 commits:
August 14, 13, 12, 8, 7, 5, 4, 1 and earlier.
:
Key management system for Zama protocol.
Last commit 5 hours ago:
There are 1,496 of them:
14, 13, 12, 11, 8, 7, 6, 5, 4, 1 August and earlier.
:
Dapp SDK for FHEVM protocol.
Last commit 2 hours ago:
Total - 565 commits:
14, August 1 and earlier.
:
A carefully curated list of awesome resources for fully homomorphic encryption (FHE), created by the Zama team.
Last commit last week:
Total 149 commits:
August 6, 31, 2, 1 July and earlier.
:
This repository contains examples of dApps created using FHEV (Fully Homomorphic EVM). Each example demonstrates various aspects of creating privacy-preserving smart contracts using FHE operations.
Last updated 3 days ago:
Total 112 commits:
11, 4 August, 22, 13, 10, 9, 8 May and earlier.
:
Updated 3 weeks ago:
Total of 312 commits:
July 25, 23, 22, 21, May 13 and earlier.
Source - the last repository researched earlier zama-ai/bounty-program.
The Zama Bounty Program is an experimental program launched by Zama to incentivize developers to participate in the development of the fully homomorphic encryption (FHE) space. The program offers monetary rewards for solving specific technical challenges that can significantly advance FHE technologies.
The goal of this initiative is to inspire and motivate developers to create applications using Zama libraries and promote FHE technologies.
The program places high value on innovation and contribution to development, rather than bug fixes or maintenance tasks.
What is expected from participants
Each bounty is assigned €10,000 in prize money:🥇 Best solution - up to €5,000🥈 Second place - up to €3,000🥉 Third place - up to €2,000.
To join the program, you need to:
Go to the repository with the list of bounties.
Register for the program through Zama Guild - an email will be sent from there with access to the portal for submitting the solution.
The code is evaluated based on several criteria:
technical efficiency and correctness of the solution,
code quality (cleanliness, structure, readability),
presence of explanatory documentation,
performance of the solution.
Program Committee
Claims are reviewed by a committee consisting of Zama team members and external experts.
They are responsible for accepting applications, evaluating solutions, accepting pull requests, and paying rewards.
Rating 4 out of 5: it's open, updates are frequent and from different developers. The only thing - not sure if there is the blockchain code itself.
There is a bounty program, but it doesn't cover vulnerabilities, just FHE application development. No audits are mentioned - that's a minus.
The website has a page with demo examples:
Let's look at them below.
Website https://portfolio.demo.zama.ai/:
"Connect Wallet" - nothing happens.
By the way, there is a message about pausing the demo.
Upload one of your favorite photos to see how you can apply a filter and save it in encrypted form.
"Drop image here or click to upload" and I upload:
Enter or generate a password to encrypt the image (so the cloud service can't scan it) and click "ENCRYPT IMAGE":
Select a filter and "APPLY FILTERS":
Enter your password and "DECRYPT NOW":
I think it's ready!
On the second tab "Confidential Token Transfer":
Click "Reveal Balance" for both to display the balances:
The screenshot only shows the balance for the first one, as the transfer form appears immediately after revealing the balance of the second wallet:
Enter the amount 1 and click "Encrypt Amount":
"Proceed with Transfer":
Clicked "Reveal Balance" for both:
Success!
Visual example. It goes without saying that there are probably no transactions in the blockchain at all, but the visual clarity is important that balances are hidden by default.
Demo site: https://huggingface.co/spaces/zama-fhe/encrypted_credit_scoring:
Step 1: Generate keys. Registration of the applicant, Bank, and credit bureau
The private key is generated jointly by organizations that jointly calculate the credit rating. It is used for encrypting and decrypting data and will never be transferred to third parties.
The evaluation key is a public key required by the server to process encrypted data. Thus, it is also sent to the server for further processing.
"Generate the keys and send evaluation key to the server.":
Step 2: Fill in some information.
Settings for the applicant, bank, and credit bureau
Select the information corresponding to the profile you want to evaluate. This model presents three sources of information:
applicant's personal data necessary for assessing their compliance with credit card requirements.
The applicant's bank account history, which contains any information about the applicant's bank details related to decision-making (here we consider the account's validity period);
and the credit bureau's information, which is any other information (in this case, the work book) that can provide additional information needed for decision-making.
Please always encrypt and send the values (using the buttons on the right) after updating, before running FHE output.
Completed step 2.1. Left a lot as is:
"Encrypt the inputs and send to server.":
In the field below, we see what was encrypted.
In step 2.2, I also did "Encrypt the inputs and send to server.":
Step 2.3:
Slightly changed - version with the number of years. "Encrypt the inputs and send to server."
Step 3: Run the FHE evaluation.
Server side
Once the server receives the encrypted input data, it can compute the prediction without needing to decrypt any values.
This server uses a decision tree classifier model that was trained on a synthetic dataset.
"Run the FHE evaluation."
Done.
Step 4: Get the encrypted output data from the server and decrypt them.
Decryption by applicant, bank, and credit bureau
Once the server finishes the output, the encrypted output data will be returned to the applicant.
Only three organizations providing information for calculating the credit rating can decrypt the result. They participate in the decryption protocol, which allows decrypting the full result only when all three parties decrypt their share of the result.
The first value shown below
is the abbreviated byte representation of the actual encrypted output data. Then the applicant can decrypt the value using their private key.
"Receive the encrypted output from the server.":
We see the value and "Credit card, most likely to be approved ✅".
Step 5: Explain the forecast (only if the credit card is most likely to be denied in the output).
If the credit card is most likely to be denied, the applicant can ask how many years of employment, most likely, will be required to increase the likelihood of credit card approval.
For simplicity, all the above actions are combined into one button. Thus, the next button encrypts the same data (except for the employment history, which varies) from all three parties and launches a new forecast
in FHE and decrypts the output data.
If the instructions below suggest entering a new input "Years of Employment", you can simply update the value at step 2 and directly perform step 6 again.
Encrypt the input data, perform calculations in FHE and decrypt the output data.
To run "Encrypt the inputs, compute in FHE and decrypt the output." I didn't need to change anything. I refreshed the page, left everything as default and clicked again:
Here's what it says:
"Your credit card application will most likely be denied ❌ However, having work experience of at least 2-5 years will increase your chances of getting your credit card approved."
Site https://huggingface.co/spaces/zama-fhe/encrypted_health_prediction:
Step 1: Select main complaints
On the client side
Select at least 5 main complaints from the list below.
Digestive system problems▼
Urological problems▼
Vascular and Lymphatic System Issues▼
Acute Respiratory Viral Infections (ARVI) Issues▼
Dermatological Issues▼
Musculoskeletal System Issues▼
Ophthalmic Issues▼
Chest Issues▼
General Issues▼
For example, I will select "General Concerns▼":
Also in other sections.
Sent after clicking "Submit":
Step 2: Data Encryption
On the client side
Key generation
In FHE schemes for encrypting and decrypting client-owned data, private encryption/decryption keys are generated.
Additionally, a public evaluation key is generated, allowing external entities to perform homomorphic operations on encrypted data without needing to decrypt it.
The evaluation key will be sent to the server for further processing.
"Generate the private and evaluation keys."
"Encrypt the data using the private secret key"
"Send data"
Step 3: Run the FHE evaluation
Server-side
Once the server receives the encrypted data, it can process and compute the output without decrypting the data, as it would have been with open data.
This server uses a logistic regression model that was trained on this dataset.
"Run the FHE evaluation":
Step 4: Decrypt the data
On the client side
Get the encrypted data from the server side
"Get data":
"Decrypt the output using the private secret key", and you will get the English versions of the diseases based on your selection:
Well, something like that. I'll repeat after them that this is a demo, not real diagnoses.
Site https://huggingface.co/spaces/zama-fhe/encrypted_dna:
Step 1: Create a genetic ancestry simulator
To run this demo, the simulator randomly selects N individuals from G generations from a set of genetic data.
Each individual is represented by two alleles from chromosome 22, which is especially important for tracking human evolutionary history and migration models around the world. By analyzing specific markers known as single nucleotide polymorphisms (SNPs) at key positions on this chromosome, valuable information can be obtained.
Each ancestor passes on half
of their genetic material 50% of our genes are inherited from the mother, and 50% from the father. These genes are organized in pairs of alleles, one from each parent.
A gene can be dominant, meaning it is expressed and determines the visible trait, or recessive, remaining unexpressed but still present in the genome. A recessive allele can manifest in future generations if it is passed on by both parents.
This simulation will use five different genetic populations: American, African, European, East Asian, and South Asian.
"Generate a random genetic allele":
Next, I completed step 2 with encryption by clicking the "Generate the secret and public keys" and "Encrypt the data using the secret key" buttons, and then sent the data by clicking "Send data to the server":
On step 3, I launched FHE computation on the server side by clicking the "Run FHE on the server" button and waited for 5 minutes due to the large amount of data:
"Send data to the client" and "Decrypt the data using the secret key" for decryption:
Also an interesting demonstration.
Anonymization is the process of removing personal information (PII) from a document to protect the privacy of an individual.
Encrypted anonymization uses fully homomorphic encryption (FHE) to anonymize personal information (PII) in encrypted documents, allowing computations to be performed on encrypted data.
In the example above, we show how to use encrypted anonymization to use LLM services like ChatGPT while preserving privacy.
Page https://huggingface.co/spaces/zama-fhe/encrypted-anonymization:
Generating keys by "Generate the secret and evaluation keys", left all items checked and clicked "Encrypt the document":
Left the query as is and clicked "Encrypt the prompt":
"Anonymize using FHE" - anonymize with FHE:
Sending the anonymized prompt to Chat GPT via "Query ChatGPT":
It gives an error... Apparently they no longer have the ability to send to GPT.
Currently, you can get them on the Guild project, by completing various tasks. For example, holding at least 0.1 ETH in Ethereum mainnet.
There are options for developers and content creators.
Rating 3 out of 5: the demos are interesting, but the portfolio and one more are not fully functional. There is no application where Zama can already be used with different contracts in the testnet.
20 out of 25 points:
Team: 5 out of 5: there are many team members, and the main ones (which I analyzed) are professionals. Social media is active (there is even Youtube with Reddit) and activity is good.
They answer questions in Discord and do so correctly.
Concept: 4 out of 5: everything is described in detail. There is information about the reasons for the demand for Zama (finance, identification and management cannot be public, as in the web2 world, data is not visible to everyone. Accordingly, the success of applications on Zama is yet to come), and technical details.
The idea is also unique: a level for inter-blockchain privacy. But the downside is that the network is partially centralized and is unlikely to provide good scalability compared to existing blockchains without centralization by switching to ASICs.
Coin: 4 out of 5: the token has utility, but there is no initial distribution yet. Investments are excellent (130 million from well-known investors, such as Pantera Capital, Multicoin Capital and Anatoly Yakovenko).
Code: 4 out of 5: it is open, updates are frequent and from different developers. The only thing is that I'm not sure if there is the blockchain code itself.
There is a bounty program, but it does not cover vulnerabilities, but the development of FHE applications. Audits are not indicated - this is a minus.
Practice: 3 out of 5: the demos are interesting, but the portfolio and one more are not fully functional. There is no application where Zama can already be used with different contracts in the testnet.
Good result: better than many.
Subscribe to https://t.me/blind_dev - there are my posts with project reviews, opinions on the prospects of different projects, tokenomics analysis, and news about my developments.
I would be very happy to spread the review article and donations.
Zama makes familiar blockchains private: now smart contracts can process encrypted data that is not even visible to nodes. This is a level of privacy without migrating to a new chain - Zama works on top of any L1 and L2. Demo applications are already available, including a confidential portfolio and other use cases.
Let's explore the project: how good is it.
There is information about the team in the "About" section and in more detail (albeit without LinkedIn links) in the Litepaper:
Rand Hindi, CEO of Zama. Work experience: Co-founder of Zama since 2020 to present; was a partner at Unit.vc; founded the startup Snips, which was acquired by Sonos; previously - member of the French Digital Council (advised the government on AI and Privacy issues); was a lecturer at Science Po University in Paris; advisor to several companies in biotech, AI and defense.
Education: BSc Computer Science and PhD Bioinformatics, University College London UCL.
Pascal Paillier, CTO of Zama. Work experience: co-founder of Zama since 2020; inventor of the Paillier additive homomorphic encryption scheme; before Zama, he headed the cryptographic innovations team at Gemalto; founded the consulting firm CryptoExperts; 2025 - IACR Fellow; led several ISO cryptography standards.
Education: PhD in Cryptography, Telecom Paris.
Nigel Smart, Chief Academic Officer of Zama. Work experience: one of the leading researchers and executives of the company; specific workplaces and dates are not specified.
Marc Joye, Chief Scientist of Zama. Work experience: one of the leading researchers and executives of the company; specific workplaces and dates are not specified.
Also found on the site LinkedIn:
10 thousand followers.
Last post a week ago. It received 46 reactions and 4 reposts. They wrote about the story of Zama and FHE.
As many as 170 participants associated with the company!
I think this is too many... But I may not know the specifics of developing such projects.
I selected a small list of key employees for analysis.
8,791 followers.
Shares other content in posts, but also about Zama.
Work experience
Co-founder & CEO at Zama since November 2019
General Partner at Unit Ventures since January 2020
Advisor at CrunchDAO since February 2022
Advisor at Mindstate Design Labs since June 2021
Advisor at Chipiron since February 2021
Advisor at NAVAL GROUP from September 2018 to December 2021
Lecturer at Sciences Po from September 2017 to December 2021
Advisor at Sonos, Inc. from January 2020 to January 2021
Total 14.
Education
At UCL: PhD, Bioinformatics from 2006 to 2011
At UCL: Bsc, Computer Science from 2003 to 2006
THNK School of Leadership: Graduate Degree, Creative Leadership from 2012 to 2013
Singularity University: Graduate program, Exponential Technologies in 2010
Good experience.
4,920 followers
No publications.
Work experience
Chief Technology Officer at Zama since November 2019
At CryptoExperts: Chief Executive Officer from February 2009 to December 2019; Senior Cryptography Expert from February 2009.
Non Executive Advisor at Nitromia since May 2019
Non Executive Advisor at CircaGene since April 2019
Non Executive Advisor at Olvid since March 2019
Co-editor of ISO/IEC 27551 "Requirements for attribute-based unlinkable entity authentication" at ISO since April 2017
Editor of ISO/IEC 18033-6 "Homomorphic Encryption" at ISO from October 2015 to October 2019
Total 11.
Education:
Telecom Paris: PhD, Computer Science, Cryptography from 1995 to 1999
Extensive experience. Great that he has worked in leadership positions and engaged in cryptography + FHE - this is useful for the project.
10,020 followers.
Last post 1 day ago with thanks for 10K followers with 54 reactions and 7 comments.
Previous - 6 days ago with 6 reactions. Post about tGBP integration, "Wirex."
Work experience:
Director of Business Development at Zama since February 2024
Membre at Crypto Valley Association since March 2024
Mentor at Outlier Ventures since May 2024
At Ledger: Inside Sales - Ledger Enterprise from January 2020 to December 2020; Sales Director - EMEA - Ledger Enterprise from January 2021 to December 2021; Sales Director - UKI & Nordics - Ledger Enterprise from January 2022 to January 2024
Sales Development Representative at Snips from 2019 to 2020
Business Developer at Médoucine from 2017 to 2019
Commercial Development / Pre-sales at Worldline in 2016
Education
Frankfurt School of Finance & Management: DeFi Talents, Decentralized Finance
Conservatoire National des Arts et Métiers: Fundamentals of Blockchain Technology, Finance, General
IÉSEG School of Management: MSc in Management - Grande école, International Business, Economics & Strategy
Good work experience and profile. Experience in Ledger Enterprise directly supports the current BizDev role.
4,023 followers.
Last post 1 month ago about Zama with 32 reactions and 3 reposts.
Work Experience
Ecosystem Lead at Zama since January 2023
At Theodo: Product Manager from March 2017 to March 2020; Head of blockchain BU and nocode BU from March 2020 to March 2022
Co-founder at Bannann from September 2015 to June 2018
Equity Derivative Sales at BNP Paribas from January 2014 to August 2015
Investor Relations Assistant at Anaxago from February 2013 to August 2013
Education
ESCP Business School: Master in Management Grande Ecole from 2011 to 2015
Queen’s University: 6-Months Exchange Program, Business / Commerce in 2014
Lycée Kléber: Classe préparatoire aux grandes écoles - Economique et commerciale option Scientifique (ECS) from 2009 to 2011
He has experience in product management and building blockchain ecosystems - this is useful for the current role of Ecosystem Lead.
2,272 followers.
Last post 2 months ago with 61 reactions. About Zama and privacy blockchains.
Work Experience:
VP - Corporate Development at Zama since October 2022
Strategy Management Consultant from April 2022 to October 2022
At Digital14: VP - PMO and Corporate Development from August 2015 to April 2017; SVP - Strategy, Performance Management and Corporate Development from April 2017 to October 2019; Chief Financial Officer and Head of Strategy from October 2019 to December 2021
Strategy Management Consultant from August 2014 to August 2015
CEO and Co-Founder at KooraBazar.com from January 2013 to July 2014
VP - Investment and Asset Management in ICT at Mubadala Development Company from January 2009 to January 2013
Engagement Manager at Booz & Company from November 2006 to December 2008
Line Manager at Axalto from July 2003 to July 2005
Total 11.
Education
INSEAD: MBA, Finance / Entrepreneurship / Strategy from August 2005 to July 2006
Stanford University: Master of Science (MS), Electrical and Electronics Engineering from September 1998 to June 2000
École Polytechnique: Diplome de l'Ecole Polytechnique, Maths / Electronics from September 1995 to June 1998
Lycee Prive Ste Genevieve: Maths from September 1993 to June 1995
Тоже хороший опыт у него.
3 973 отслеживающих.
Последняя публикация 1 неделю назад с 38 реакциями, 3 комментариями и 1 репостом. Про найм в Zama.
Опыт работы
В Zama: Marketing Director с октября 2021 по настоящее время; Vice President Marketing с апреля 2024 по настоящее время
Co-founder - Growth and acquisition в Siliconwood с ноября 2017 по сентябрь 2021
Ruby on Rails Developer (Freelance) с мая 2012 по декабрь 2020
Growth Hacker в Leetchi.com с августа 2014 по февраль 2017
Education
San Francisco State University: Digital Marketing and Entrepreneurship from December 2013 to July 2014
International School of Business, Dublin: Bachelor of Business Administration (B.B.A.) in 2012
ISTEC - Ecole Supérieure de Commerce et de Marketing: Master’s degree from 2010 to 2015
Ecole Suger: Baccalauréat scientifique in 2010
Longtime at Zama: held several marketing positions - a professional in this. And the project is popular, which proves it.
He has years of experience in marketing/growth (Growth Hacker, co-founder in growth, specialized training), so past roles confirm professionalism in marketing.
172 followers.
Last post 6 months ago with 10 reactions, 3 comments, and 3 reposts.
Work experience:
Senior Developer Relations at Zama from October 2024 to present
Technical co-founder at Lastic from September 2023 to present
Community Engagement Coordinator & Web Developer (Volunteer) at SubWork - coworking in nature from November 2022 to present
Co-Founder at Women Of Polkadot from October 2023 to December 2024
Founder at PoppySeed Dev from January 2023 to December 2024
- Total 15.
Education
Polkadot Blockchain Academy: Bootcamp, Blockchain Development from July 2023 to August 2023
DappCamp: Bootcamp, Blockchain Development from October 2022 to November 2022
University of Ljubljana, Faculty of Computer and Information Science: Master’s degree, Intermedia/Multimedia from October 2020 to February 2021
University of Ljubljana, Faculty of Electrical Engineering: Bachelor’s degree, Electrical and Electronics Engineering from October 2018 to July 2020
University of Ljubljana, Faculty of Mathematics and Physics: from 2016 to 2018
Good experience and education. Judging by the previous experience, may perform well in the current position at Zama.
425 followers.
Last post 1 week ago with 38 reactions, 3 comments, and 1 repost. About a community lead vacancy at Zama.
Previous post 3 weeks ago with 122 reactions, 6 comments, and 13 reposts. About Zama and OpenZeppelin partnership on confidential smart contracts.
Work experience
At Zama: Senior GPU Researcher from February 2022 to September 2022; Senior GPU Engineer from February 2022 to present
At Universidade Estadual de Campinas: Scientific initiation researcher from March 2010 to December 2011; Master of Science from February 2014 to June 2016; PhD candidate from February 2016 to March 2023
Visiting PhD Student at Aarhus University from September 2020 to September 2021
Business Analyst at Kanui from April 2013 to March 2014
Scientific initiation researcher at Nacional Laboratory of Biosciences LNBio/ABTLuS from March 2012 to July 2013
Education
Universidade Estadual de Campinas: Doctor of Philosophy (PhD), Computer Science from 2016 to 2022
Universidade Estadual de Campinas: Master of Science (MSc), Computer Science from 2014 to 2016
Universidade Estadual de Campinas: Bachelor's degree, Applied Mathematics from 2008 to 2013
Strong research and engineering background: applied mathematics → MSc/PhD in computer science, experience with GPU/parallel computing and academic research; has been at Zama for a long time (from researcher to senior engineer), indicating accumulated expertise and good relevance to the company's tasks.
99 followers.
Last post 3 weeks ago with 122 reactions, 6 comments, 13 reposts. About Zama and OpenZeppelin partnership on confidential smart contracts.
Previous post 1 month ago. Received 155 reactions, 7 comments and 26 reposts. About the announcement of Zama Protocol and the launch of the public testnet.
Work experience
Senior Solidity Engineer at Zama from April 2025 to present
Senior Lead Software Engineer, freelance from June 2021 to April 2025
Career break from 2020 to 2021
Co-Founder, Principal Software Engineer at GlobFX Technologies from 2012 to 2019
Co-Founder, Principal Software Engineer at GlobFX Technologies from 2000 to 2019
Co-Founder, Principal Software Engineer at GlobFX Technologies from 2004 to 2009
Java Software Developer at Dassault Systèmes from 1998 to 2000
Education
Grenoble INP Ensimag: Master’s degree, Computational and Applied Mathematics from 1993 to 1996
Lycée Henri IV: MPSI MP from 1992 to 1993
Good experience in development (engineering). Worked in this direction not only at Zama, but also earlier.
511 followers.
No publications.
Work Experience
Cryptographe at Zama from October 2020 to present
Chercheur en cybersécurité at WALLIX Group from December 2018 to October 2020
Doctorant sur le sujet "Evaluation de la confiance dans les architectures de sécurité" at Université Grenoble Alpes from October 2014 to August 2018
Enseignant (moniteur/ATER) at Université Grenoble Alpes from September 2015 to June 2018
Internship "Classification of semi-bodies and substitution tables for symmetric ciphers" at Laboratoire Jean Kuntzmann from February 2014 to July 2014
Industrial Project "Acceleration of algorithms for digital printing" at Caldera from January 2013 to July 2013
Education
Université Grenoble Alpes: Doctorate, Mathematics and Computer Science 2018
Grenoble INP Ensimag: Specialization in Security, Cryptography and Information Coding 2014
Grenoble INP Esisar: Engineering Degree, Computer Science and Networks 2014
Lycée Pierre d'Aragon: Scientific Baccalaureate with Mathematics specialization 2009
Professional: foundation in cryptography and security, strong skills in formal proofs and protocol design, focus on verifiability and practical applicability of solutions.
In the blog the latest post on July 30 - announcement of partnership:
Developer forum active:
Last message 16 minutes ago (August 14 at 16:44 MSK).
Active chat (wrote a minute ago). Questions are answered.
And correctly.
At first I thought they weren't getting through, but now I've reached the community section on the website, and there's a link to LinkedIn.
In announcements, the last message was on August 12 (normal):
In the creators' program chat, they also write actively:
In technical support as well:
In the developers' program chat as well:
Overall, Discord is super active!
X @zama_fhe:
215.9K subscribers.
1,699 posts. Last one on August 8 about community update (developer program winners). It got 106 replies, 115 reposts, 439 likes, and 33,427 views.
Previous one was on August 5 about meeting project founders to discuss HFE and Zama. It got 80 replies, 108 reposts, 438 likes, and 18,553 views.
X analysis by Tweetscout:
Score 771, level 3 (Credible).
Subscribed projects: @MilkRoadDaily, @flipsidecrypto, @OpenZeppelin, @protocollabs, @JSCCapital, @GSR_io, @ShytoshiKusama, @conduitxyz and others.
Venture capital: @tarunchitra, Robot Ventures, @jbrukh, CoinFund, @VinnyLingham, Newtown Partners and others.
Analysis of X by Moni:
Moni ScoreLevel: 4. Established
3551
0 per day.
136 mentions by smarties, 290 smarties.
According to Moni, the following projects are subscribed: @NEARProtocol, @Filecoin, @Shibtoken, @SecretNetwork, @OasisProtocol, @PhalaNetwork, @Try_Wink, @vita_dao and others.
VC: @robotventures, @GSR_io, @hack_vc, @Signum_Capital, @StakeCapital, @ej__rogers.
Good level of fame and activity for X.
64,107 subscribers, last post on August 2 with 8K views and > 200 reactions - that's cool!
3,93K subscribers. Last video two weeks ago with 221 views.
Last messages 1 month ago. Many reactions - > 200-300.
11 thousand subscribers. Last post 15 days ago with 49 reposts, 2 quotes, 177 likes and 2.4K views.
Previous 22 days ago with 27 reposts, 1 quote, 99 likes and 1.2K views.
Rating 5 out of 5: many team members, and the main ones (analyzed) are professionals. Social media is active (even Youtube with Reddit) and activity is good.
They answer questions in Discord and correctly.
There is documentation, and as part of it - Litepaper:
We considered the problem: blockchains are needed for finance, governance, and identification. But they are not confidential (all data is public). This is a problem. No large or medium-sized investors want everyone to see the amount of money and where it is located.
Zama is an interchain privacy layer on top of any L1/L2 that adds confidential smart contracts with programmable access rules to existing blockchains. Input data and transaction states are end-to-end encrypted, not even visible to nodes, while contracts remain composable with both private and public protocols.
Since it's not a new L1/L2, users don't need to migrate to another chain, and they can work with private dApps from anywhere.
Zama is based on fully homomorphic encryption (FHE), which allows computations to be performed directly on encrypted data without revealing it. The results remain publicly verifiable.
Zama's FHE technology is inherently post-quantum, already accelerated 100+ times, and supports familiar languages like Solidity and Python. The team calls it the next step after HTTPS - "HTTPZ", where privacy becomes the base for all on-chain applications.
For scalability and flexibility, the protocol combines FHE with MPC and ZK: keys are decentralized, inputs are efficiently verified, and performance scales with GPU and specialized hardware.
As a result, Zama creates a universal privacy layer that doesn't require changing the blockchain but is compatible with any L1 and L2.
The Zama protocol is based on years of research and development by Zama. The public testnet is already running, allowing everyone to deploy and test private dApps, and operators to set up coordination and processes. The first official mainnet is scheduled for Q4 2025 on the Ethereum blockchain: it will bring privacy to the Ethereum ecosystem, and only Zama-approved applications will be available at launch. By the end of the year, there will be a TGE with the launch of the $ZAMA token and connection of other EVM chains to the Zama protocol, after which dApp deployment will be possible without permission. In 2026, after completing the initial EVM-only phase, the protocol will be deployed in Solana, opening up confidential SVM applications.
It's great that they identified them, as it helps understand the demand scenarios for Zama.
Confidential smart contracts open up a new design space for on-chain applications, especially in finance, identification, and governance. In the web2 world, most services do not publicly disclose data, which means the lion's share of on-chain applications is yet to come thanks to Zama.
In finance, the protocol enables private payments: balances and transfer amounts are end-to-end encrypted, and compliance requirements can be embedded directly into token contracts. For tokenizing real-world assets, large institutions can use public blockchains like Ethereum and Solana while maintaining the confidentiality of operations and investor identities and performing KYC/AML in smart contracts without disclosing sensitive data. This is confirmed by a PoC in the JP Morgan - Kynexis report based on Zama technologies. In DeFi, fully encrypted swaps become possible with protection against front-running and hiding amounts and even assets, as well as private lending, on-chain credit scoring, and option pricing.
In token operations, closed auctions with sealed bids appear: participants publish encrypted bids, and only the winner is revealed at the end, improving pricing and protecting against bots from the mempool, especially in public token sales. Distributions can be conducted confidentially without disclosing amounts on the address, supporting vesting on encrypted tokens and private staking.
For identification and governance, composable on-chain identity is supported: a full DID + VC scheme lives on the blockchain in encrypted form and remains compatible with dApps, providing an abstraction of id and allowing smart contracts to verify claims for regulatory compliance privately and decentrally.
Voting becomes confidential: individual votes and stakes are hidden, only the result is revealed.
Other examples include on-chain corporations, where the capital structure, finances, board decisions, customer and employee data remain private, and smart contracts automate operational processes.
Prediction markets get encrypted forecasts with periodic disclosure, reducing biases and increasing accuracy.
Data markets for AI allow users to selectively share and sell data, train models on encrypted datasets, and receive a steady stream of income instead of a one-time sale.
This is just a part of what is already available today. As FHE scales through the Zama protocol, liquidity will grow and the transition of users and companies on-chain will become easier. Over time, it will be possible to run entire companies, cities, and even countries on the blockchain with financial and identification infrastructure, elections, currency, taxes, and registries of land, vehicles, and companies. Confidential blockchains provide not only programmable money but also programmable public infrastructure.
Most chains cannot natively perform FHE.
Zama solves this with two ideas: symbolic execution and threshold decryption.
Symbolic execution
A contract on the host chain invokes FHEVM. The chain itself does not compute FHE: it writes a pointer to the result and emits an event.
Coprocessors pick it up. They perform FHE computation and store the ciphertext.
Advantages:
no need to change the chain or install special hardware,
non-FHE transactions are not slowed down,
FHE operations can be parallelized.
Since the host chain stores pointers, not the ciphertexts themselves (data is stored by coprocessors), FHE calls are easily composable without waiting for previous ones to complete (waiting is only needed at the decryption moment).
Security:
From a security standpoint, all coprocessor actions are publicly verifiable: anyone can recalculate the ciphertexts and verify their correctness. At launch, several coprocessors with majority consensus are used, and in the future, it is planned to allow anyone to perform FHE operations with proof of correctness via ZK-FHE.
Threshold Decryption
To maintain on-chain composability, all ciphertexts are encrypted under the same public key, meaning the private key must be protected from unauthorized decryption. In the Zama protocol, the private key is split between participants and managed by a dedicated threshold MPC as a key management service (KMS). A user or contract can only decrypt a value if the source contract explicitly granted such permission on the host chain; a decryption request is sent to the Zama Gateway, which orchestrates the process and redirects it to KMS participants. All such requests are publicly visible, so anyone can verify their compliance with the access control logic defined by smart contracts.
Components
The ecosystem includes several basic components.
Host chains are L1/L2 protocols supported by Zama, on which developers deploy confidential dApps.
The FHEVM library provides them with tools to create private smart contracts, and the FHEVM Executor contract on the host chain accepts dApp calls, records FHE operations, and automatically emits events for coprocessors.
An access control list (ACL) is deployed on each host chain and tracks who is authorized to compute over an encrypted value and who can decrypt it; coprocessors relay events from the ACL to the Gateway, where ACLs from different host chains are aggregated into a single registry for authenticating KMS decryption requests.
The native $ZAMA token is used to pay fees, stake, and govern.
Gateway is a set of smart contracts that orchestrate the protocol: users send encrypted inputs for verification, request decryption, and bridge encrypted assets between host chains through it, paying a small fee in $ZAMA.
For maximum performance and efficiency, a dedicated rollup on Arbitrum is chosen: it serves only the Zama protocol and does not allow deployment of third-party contracts.
Coprocessors are nodes that verify encrypted inputs, perform FHE computations, store resulting ciphertexts, and relay ACL events to the Gateway; each node commits its result, after which the Gateway makes a decision by majority. All their tasks are publicly verifiable, and scaling is possible both vertically and horizontally.
KMS nodes run MPC protocols for key generation, CRS, and threshold decryption; no single party ever has access to the keys. KMS nodes are orchestrated via Gateway and operate within AWS Nitro Enclaves, which complicates key share leakage and provides software version attestation; ZK-MPC is planned to be added for verifiability without relying on hardware.
Operators are organizations that run protocol nodes, including coprocessors and KMS.
Performance
The Zama protocol was initially designed to be horizontally scalable and relies on the TFHE-rs library. Unlike the sequential EVM model, FHE operations are computed in parallel, and if a specific ciphertext is not involved in a sequential chain, coprocessors increase throughput by simply adding servers.
Since the start of protocol development, throughput has increased from 0.5 to 20+ transactions per second per host chain, which in total across all host chains theoretically provides hundreds of tps. This is sufficient for most EVM chains (Ethereum ~15 tps), but insufficient for retail payments (VISA ~25,000 tps) and for Solana (1000+ tps).
The gap is planned to be closed gradually through hardware: transitioning from CPU to GPU will provide ~50-100 tps per chain (for state-independent transactions), an open FPGA accelerator aims for 500-1000 tps per chain, and specialized ASICs - for 10,000+ tps per chain.
The key point is that FHE limits are now primarily defined by 'Moore's Law' - the better the hardware, the higher the protocol's throughput.
Security
The approach is defense-in-depth with a combination of multiple layers of protection. All FHE operations use 128-bit cryptographic strength and a p-fail probability of 2^-128; the scheme is post-quantum, meaning it is resistant to quantum attacks. All FHE computations are publicly verifiable: anyone can recalculate the results and detect dishonest FHE nodes; this is similar to the security of optimistic rollups, but applied to FHE. At the same time, the results are signed not by one, but by three independent FHE nodes from different operators, providing both 'optimism' and consensus at the output.
The MPC protocol involves 13 nodes with a 2/3 majority rule, and the protocol itself works and is correct even with up to 1/3 malicious participants - according to the team, this is the first production implementation of a working MPC.
MPC runs in AWS Nitro Enclaves, adding a layer of protection and preventing access to the share of the private FHE key from outside; software attestation allows the protocol to track updates. The combination of MPC and Nitro means that compromising keys would require collusion between AWS and several MPC nodes.
Genesis operators are public and reputable organizations with billions at stake in their core business (validators, infrastructure providers, etc.); being doxxed, they bear reputational and economic risks for any violation.
Slashing is managed through governance, allowing for proposing sanctions and addressing edge cases specifically.
The protocol is undergoing one of the largest audits in the industry by Trail of Bits and Zenith - totaling over 34 audit weeks.
Compliance
Compliance rules are defined in smart contracts.
The protocol does not decide who can decrypt what, but the application developers do.
Development plans
Achieve 10,000+ tps:
new FHE techniques with 10-20x acceleration,
FHE ASIC with 100-1000x acceleration,
migrate Gateway to ZK rollup with less than 100 ms delay.
Strengthen KMS:
ZK-MPC for trustless correctness verification,
more nodes in the MPC committee - dozens and hundreds.
Open participation to operators:
launch MPC inside HSM,
ZK-FHE, where correctness is proven, and coprocessors compete like in PoW.
Currently expensive, but progress is being made.
Full post-quantum stack:
Replace ZKPoK with lattice-based post-quantum, host-chain signatures are not post-quantum yet.
L1/L2 community migration.
The Zama protocol uses Delegated Proof-of-Stake.
It is run by 16 operators - initially 13 KMS nodes and 3 FHE coprocessors (the number will grow over time).
Operators are selected according to the following rules:
Genesis operators are selected based on reputation, DevOps experience, and offchain value (equity, revenue, market cap, etc.). This provides initial security through reputation.
A large operator risks losing customers if caught violating the Zama protocol.
Eventually, anyone will be able to become a KMS operator or coprocessor. You need to reliably demonstrate node operation in the testnet, and then stake at least 0.5% of the circulating supply of $ZAMA. Each epoch (initially 3 months), the protocol selects the top 13 KMS and top 3 coprocessors by stake for the next epoch.
All 16 active operators receive rewards in $ZAMA. The size depends on the role and stake share.
The team is currently looking for genesis operators.
If you want to apply - fill out the form.
Token holders without the necessary infrastructure can participate through Delegation.
They delegate their $ZAMA to operators from the whitelist and receive a portion of their rewards.
Each operator independently chooses the motivation for delegators - reduced commission or additional rewards not in $ZAMA.
Most operators adopt protocol updates. This includes software updates, commission changes, adding new host chains, etc.
Exception - emergency protocol pause and spammer blacklist, any operator can do this.
Unpausing or removing from the blacklist can only be done through regular social consensus.
Slashing is provided for abuse. This way the protocol quickly responds to critical cases and encourages honest behavior.
On-chain voting by $ZAMA holders is used for two things:
changing inflation,
slashing operators who break the rules.
There is an official dashboard:
Transaction fees (total) - 1,082 Sepolia ETH
529,046 transactions
108,822 users
6,979 contracts
Rating 4 out of 5: everything is described in detail. There is information about the reasons for the demand for Zama (finance, identification, and management cannot be public, as in the web2 world, not everyone can see the data. Accordingly, the success of applications on Zama is yet to come), and technical details.
The idea is also unique: a level for inter-blockchain privacy. But the downside is that the network is currently partially centralized and is unlikely to provide good scalability compared to existing blockchains without centralization by transitioning to ASICs.
The litepaper also contains information about the token
$ZAMA is the native token of the protocol. It is needed for fees, staking, and governance.
The burn-and-mint model is used - 100% of fees are burned, new tokens are minted for operator rewards.
Fee model
Deploying a confidential application on a supported network is free and permissionless. The Zama protocol does not charge a fee for FHE calculations.
Payment is charged for three actions:
ZKPoK verification. The user pays when adding encrypted inputs to a transaction. The price depends on the number of bits being verified.
Decryption of ciphertexts. The user pays for each decryption request. The price depends on the number of bits to be decrypted.
Bridging of ciphertexts. The user pays for transferring an encrypted value between networks. The price depends on the number of bridged bits.
The commission can be paid by the end user, frontend, or relayer.
Developers can make it so that users don't need to hold $ZAMA.
Fees are paid in $ZAMA but denominated in USD.
The oracle regularly updates the $ZAMA/USD price on the Gateway.
Gateway contracts recalculate how much $ZAMA is needed for each function.
This creates two effects:
fees are tied to usage, not speculation,
it's easier for users, developers, and relayers to plan expenses in USD.
Volume discount applies:
The more an address verifies/decrypts/bridges in 30 days, the lower the price per bit.
Discounts range from 10% to 99% depending on volume.
Starting price grid:
ZKPoK verification: $0.016 to $0.0002 per bit,
Bridging: $0.016 to $0.0002 per bit,
Decryption: $0.0016 to $0.00002 per bit.
Decryption is 10 times cheaper than ZKPoK and bridging. This is due to frequency - a 10:1 ratio in favor of decryptions is expected.
Example of confidential token transfer:
amounts and balances take up 64 bits,
usually 3 decryptions are needed per transaction - sender's balance, recipient's balance, and final amount (0 if transfer failed),
then the final cost, including discounts:
ZKPoK verification: 64 * ZPoK_bit_price = $0.01 to $1,
decryption of 2 balances + amount: 64 * 3 * decryption_bit_price = $0.003 to $0.30,
total: about $0.01 to $1.3 on average.
The model makes operations accessible to large users and profitable for operators.
A small occasional user will pay about $1 per transaction. A large user (e.g., wallet), on the other hand, will pay up to $0.01 per transaction.
Base prices are set so that 1 tps on a single host chain yields about $3.5m in fees per year.
If 10% of global blockchain transactions use Zama for privacy (about 300 tps), the protocol's annual revenue could reach $1 BILLION.
Staking rewards
Operators stake $ZAMA to participate in the protocol and receive rewards. They are minted by inflation - initially 10%.
The rate can be changed on-chain.
Distribution is by roles - sequencer, coprocessors, KMS. Then rewards are divided proportionally to stake within the group.
The operator decides how to share with delegators - expected commission of about 20%.
Expected parameters by groups:
Coprocessors - 3 operators - about 13.3% rewards per operator. Cost estimate: $20,000 per month for every 10 tps host chains.
KMS - 13 operators - about 4.6% rewards per operator. Cost estimate: $2,500 per month for 50 tps decryptions.
This scheme rewards contributions based on actual work done. It also motivates delegators to carefully choose who to delegate to.
Distribution
No details yet.
According to Cryptorank, raised $130 MLN with a valuation of $1 MLRD!:
Investors: Pantera Capital, Multicoin Capital, Anatoly Yakovenko, Protocol Labs, Blockchange Ventures, Stake Capital Group, Metaplanet, Vsquared Ventures, Juan Benet and Gavin Wood.
Rating 4 out of 5: token utility exists, but no initial distribution yet. Investments are excellent (130 MLN $ from well-known investors, e.g., Pantera Capital, Multicoin Capital and Anatoly Yakovenko).
There are GitHub repositories:
There are 64 of them, so we'll look at the main ones.
:
FHEV, a full-featured framework for integrating fully homomorphic encryption (FHE) with blockchain applications.
Last commit 44 minutes ago:
Total 3,245 commits:
August 14, 13, 12, 11, 8, 7, 6, 5, 4, 3, 1 and earlier.
:
TFHE-rs: Pure Rust implementation of TFHE scheme for logical and integer arithmetic on encrypted data.
Last commit 6 hours ago:
Total 3,192 commits:
August 14, 13, 12, 8, 7, 5, 4, 1 and earlier.
:
Key management system for Zama protocol.
Last commit 5 hours ago:
There are 1,496 of them:
14, 13, 12, 11, 8, 7, 6, 5, 4, 1 August and earlier.
:
Dapp SDK for FHEVM protocol.
Last commit 2 hours ago:
Total - 565 commits:
14, August 1 and earlier.
:
A carefully curated list of awesome resources for fully homomorphic encryption (FHE), created by the Zama team.
Last commit last week:
Total 149 commits:
August 6, 31, 2, 1 July and earlier.
:
This repository contains examples of dApps created using FHEV (Fully Homomorphic EVM). Each example demonstrates various aspects of creating privacy-preserving smart contracts using FHE operations.
Last updated 3 days ago:
Total 112 commits:
11, 4 August, 22, 13, 10, 9, 8 May and earlier.
:
Updated 3 weeks ago:
Total of 312 commits:
July 25, 23, 22, 21, May 13 and earlier.
Source - the last repository researched earlier zama-ai/bounty-program.
The Zama Bounty Program is an experimental program launched by Zama to incentivize developers to participate in the development of the fully homomorphic encryption (FHE) space. The program offers monetary rewards for solving specific technical challenges that can significantly advance FHE technologies.
The goal of this initiative is to inspire and motivate developers to create applications using Zama libraries and promote FHE technologies.
The program places high value on innovation and contribution to development, rather than bug fixes or maintenance tasks.
What is expected from participants
Each bounty is assigned €10,000 in prize money:🥇 Best solution - up to €5,000🥈 Second place - up to €3,000🥉 Third place - up to €2,000.
To join the program, you need to:
Go to the repository with the list of bounties.
Register for the program through Zama Guild - an email will be sent from there with access to the portal for submitting the solution.
The code is evaluated based on several criteria:
technical efficiency and correctness of the solution,
code quality (cleanliness, structure, readability),
presence of explanatory documentation,
performance of the solution.
Program Committee
Claims are reviewed by a committee consisting of Zama team members and external experts.
They are responsible for accepting applications, evaluating solutions, accepting pull requests, and paying rewards.
Rating 4 out of 5: it's open, updates are frequent and from different developers. The only thing - not sure if there is the blockchain code itself.
There is a bounty program, but it doesn't cover vulnerabilities, just FHE application development. No audits are mentioned - that's a minus.
The website has a page with demo examples:
Let's look at them below.
Website https://portfolio.demo.zama.ai/:
"Connect Wallet" - nothing happens.
By the way, there is a message about pausing the demo.
Upload one of your favorite photos to see how you can apply a filter and save it in encrypted form.
"Drop image here or click to upload" and I upload:
Enter or generate a password to encrypt the image (so the cloud service can't scan it) and click "ENCRYPT IMAGE":
Select a filter and "APPLY FILTERS":
Enter your password and "DECRYPT NOW":
I think it's ready!
On the second tab "Confidential Token Transfer":
Click "Reveal Balance" for both to display the balances:
The screenshot only shows the balance for the first one, as the transfer form appears immediately after revealing the balance of the second wallet:
Enter the amount 1 and click "Encrypt Amount":
"Proceed with Transfer":
Clicked "Reveal Balance" for both:
Success!
Visual example. It goes without saying that there are probably no transactions in the blockchain at all, but the visual clarity is important that balances are hidden by default.
Demo site: https://huggingface.co/spaces/zama-fhe/encrypted_credit_scoring:
Step 1: Generate keys. Registration of the applicant, Bank, and credit bureau
The private key is generated jointly by organizations that jointly calculate the credit rating. It is used for encrypting and decrypting data and will never be transferred to third parties.
The evaluation key is a public key required by the server to process encrypted data. Thus, it is also sent to the server for further processing.
"Generate the keys and send evaluation key to the server.":
Step 2: Fill in some information.
Settings for the applicant, bank, and credit bureau
Select the information corresponding to the profile you want to evaluate. This model presents three sources of information:
applicant's personal data necessary for assessing their compliance with credit card requirements.
The applicant's bank account history, which contains any information about the applicant's bank details related to decision-making (here we consider the account's validity period);
and the credit bureau's information, which is any other information (in this case, the work book) that can provide additional information needed for decision-making.
Please always encrypt and send the values (using the buttons on the right) after updating, before running FHE output.
Completed step 2.1. Left a lot as is:
"Encrypt the inputs and send to server.":
In the field below, we see what was encrypted.
In step 2.2, I also did "Encrypt the inputs and send to server.":
Step 2.3:
Slightly changed - version with the number of years. "Encrypt the inputs and send to server."
Step 3: Run the FHE evaluation.
Server side
Once the server receives the encrypted input data, it can compute the prediction without needing to decrypt any values.
This server uses a decision tree classifier model that was trained on a synthetic dataset.
"Run the FHE evaluation."
Done.
Step 4: Get the encrypted output data from the server and decrypt them.
Decryption by applicant, bank, and credit bureau
Once the server finishes the output, the encrypted output data will be returned to the applicant.
Only three organizations providing information for calculating the credit rating can decrypt the result. They participate in the decryption protocol, which allows decrypting the full result only when all three parties decrypt their share of the result.
The first value shown below
is the abbreviated byte representation of the actual encrypted output data. Then the applicant can decrypt the value using their private key.
"Receive the encrypted output from the server.":
We see the value and "Credit card, most likely to be approved ✅".
Step 5: Explain the forecast (only if the credit card is most likely to be denied in the output).
If the credit card is most likely to be denied, the applicant can ask how many years of employment, most likely, will be required to increase the likelihood of credit card approval.
For simplicity, all the above actions are combined into one button. Thus, the next button encrypts the same data (except for the employment history, which varies) from all three parties and launches a new forecast
in FHE and decrypts the output data.
If the instructions below suggest entering a new input "Years of Employment", you can simply update the value at step 2 and directly perform step 6 again.
Encrypt the input data, perform calculations in FHE and decrypt the output data.
To run "Encrypt the inputs, compute in FHE and decrypt the output." I didn't need to change anything. I refreshed the page, left everything as default and clicked again:
Here's what it says:
"Your credit card application will most likely be denied ❌ However, having work experience of at least 2-5 years will increase your chances of getting your credit card approved."
Site https://huggingface.co/spaces/zama-fhe/encrypted_health_prediction:
Step 1: Select main complaints
On the client side
Select at least 5 main complaints from the list below.
Digestive system problems▼
Urological problems▼
Vascular and Lymphatic System Issues▼
Acute Respiratory Viral Infections (ARVI) Issues▼
Dermatological Issues▼
Musculoskeletal System Issues▼
Ophthalmic Issues▼
Chest Issues▼
General Issues▼
For example, I will select "General Concerns▼":
Also in other sections.
Sent after clicking "Submit":
Step 2: Data Encryption
On the client side
Key generation
In FHE schemes for encrypting and decrypting client-owned data, private encryption/decryption keys are generated.
Additionally, a public evaluation key is generated, allowing external entities to perform homomorphic operations on encrypted data without needing to decrypt it.
The evaluation key will be sent to the server for further processing.
"Generate the private and evaluation keys."
"Encrypt the data using the private secret key"
"Send data"
Step 3: Run the FHE evaluation
Server-side
Once the server receives the encrypted data, it can process and compute the output without decrypting the data, as it would have been with open data.
This server uses a logistic regression model that was trained on this dataset.
"Run the FHE evaluation":
Step 4: Decrypt the data
On the client side
Get the encrypted data from the server side
"Get data":
"Decrypt the output using the private secret key", and you will get the English versions of the diseases based on your selection:
Well, something like that. I'll repeat after them that this is a demo, not real diagnoses.
Site https://huggingface.co/spaces/zama-fhe/encrypted_dna:
Step 1: Create a genetic ancestry simulator
To run this demo, the simulator randomly selects N individuals from G generations from a set of genetic data.
Each individual is represented by two alleles from chromosome 22, which is especially important for tracking human evolutionary history and migration models around the world. By analyzing specific markers known as single nucleotide polymorphisms (SNPs) at key positions on this chromosome, valuable information can be obtained.
Each ancestor passes on half
of their genetic material 50% of our genes are inherited from the mother, and 50% from the father. These genes are organized in pairs of alleles, one from each parent.
A gene can be dominant, meaning it is expressed and determines the visible trait, or recessive, remaining unexpressed but still present in the genome. A recessive allele can manifest in future generations if it is passed on by both parents.
This simulation will use five different genetic populations: American, African, European, East Asian, and South Asian.
"Generate a random genetic allele":
Next, I completed step 2 with encryption by clicking the "Generate the secret and public keys" and "Encrypt the data using the secret key" buttons, and then sent the data by clicking "Send data to the server":
On step 3, I launched FHE computation on the server side by clicking the "Run FHE on the server" button and waited for 5 minutes due to the large amount of data:
"Send data to the client" and "Decrypt the data using the secret key" for decryption:
Also an interesting demonstration.
Anonymization is the process of removing personal information (PII) from a document to protect the privacy of an individual.
Encrypted anonymization uses fully homomorphic encryption (FHE) to anonymize personal information (PII) in encrypted documents, allowing computations to be performed on encrypted data.
In the example above, we show how to use encrypted anonymization to use LLM services like ChatGPT while preserving privacy.
Page https://huggingface.co/spaces/zama-fhe/encrypted-anonymization:
Generating keys by "Generate the secret and evaluation keys", left all items checked and clicked "Encrypt the document":
Left the query as is and clicked "Encrypt the prompt":
"Anonymize using FHE" - anonymize with FHE:
Sending the anonymized prompt to Chat GPT via "Query ChatGPT":
It gives an error... Apparently they no longer have the ability to send to GPT.
Currently, you can get them on the Guild project, by completing various tasks. For example, holding at least 0.1 ETH in Ethereum mainnet.
There are options for developers and content creators.
Rating 3 out of 5: the demos are interesting, but the portfolio and one more are not fully functional. There is no application where Zama can already be used with different contracts in the testnet.
20 out of 25 points:
Team: 5 out of 5: there are many team members, and the main ones (which I analyzed) are professionals. Social media is active (there is even Youtube with Reddit) and activity is good.
They answer questions in Discord and do so correctly.
Concept: 4 out of 5: everything is described in detail. There is information about the reasons for the demand for Zama (finance, identification and management cannot be public, as in the web2 world, data is not visible to everyone. Accordingly, the success of applications on Zama is yet to come), and technical details.
The idea is also unique: a level for inter-blockchain privacy. But the downside is that the network is partially centralized and is unlikely to provide good scalability compared to existing blockchains without centralization by switching to ASICs.
Coin: 4 out of 5: the token has utility, but there is no initial distribution yet. Investments are excellent (130 million from well-known investors, such as Pantera Capital, Multicoin Capital and Anatoly Yakovenko).
Code: 4 out of 5: it is open, updates are frequent and from different developers. The only thing is that I'm not sure if there is the blockchain code itself.
There is a bounty program, but it does not cover vulnerabilities, but the development of FHE applications. Audits are not indicated - this is a minus.
Practice: 3 out of 5: the demos are interesting, but the portfolio and one more are not fully functional. There is no application where Zama can already be used with different contracts in the testnet.
Good result: better than many.
Subscribe to https://t.me/blind_dev - there are my posts with project reviews, opinions on the prospects of different projects, tokenomics analysis, and news about my developments.
I would be very happy to spread the review article and donations.
No comments yet