Anton Dziatkovskii

Systems builder · distributed systems & applied cryptography · autonomous AI agents

I build and study complex multi-actor systems. For over a decade my work has centered on one question: how decentralized systems of many independent actors stay stable and governable — first in fault-tolerant infrastructure, consensus mechanisms and applied cryptography, then in token economics and DAO governance, and now in swarms of autonomous AI agents. I apply homeostatic control theory to these systems and validate it in production: most recently a multi-agent consensus system for LLM agents, shipped with a reproducible evaluation harness. I trained in cryptography and data security at Russia’s national nuclear university — M.Sc. in Information Security, NRNU MEPhI (Obninsk Institute of Atomic Energy), with coursework in quantum cryptography and quantum information — and later completed a Ph.D. (Information Technologies: AI, zero-knowledge proofs, blockchain, data science). I treat AI engineering the way experimental science treats a noisy instrument: measure, calibrate, state your uncertainty.

Russian readers: essays in Russian are collected here. New here? Start with one page of proof →

Selected work

Autonomous AI-agent systems

claude-consensus — multi-agent consensus for LLM agents. Open source, MIT; stdlib-only; offline eval harness with self-verifying scenarios (2026)
code
Homeostatic Governance: A Gorsky–Anokhin Alternative to the VSM. Preprint, draft (2026); US provisional patent pending — three-circuit control model for multi-actor systems, with agent-swarm mapping; claude-consensus is the reference implementation
Le Chatelier Tokenomics: negative feedback as homeostasis in multi-asset coupled equilibria. Preprint, draft (2026)

Distributed systems & applied cryptography

State data security backed by artificial intelligence and zero-knowledge proofs. Economic Annals-XXI, 202 (2023)
DOI
Researching the usage of blockchain in higher education. With U. Hryneuski. IEEE TELE (2021)
IEEE Xplore
Blockchain-based framework for efficient student performance tracking (BloSPer). IJACSA 13(11) (2022)
DOI

Stability, governance & systems analysis

Chronological progress of blockchain in science, technology, engineering and math (STEM): a systematic analysis. Sustainability 14(19) (2022)
DOI
The use of blockchain technology and homeostatic principles in the monetary policy of states. With V. Grynevski (2021)
text
Research on contour feature extraction of multiple sports images based on nonlinear mechanics. With J. Wang, M. Li, U. Hryneuski, A. Krylova. Nonlinear Engineering 11 (2022)
DOI

All publications — 50+ items, 2018–2026 → · Google Scholar · ORCID

Writing

No sandwich, please! Popular DeFi attack strategy analysis. HackerNoon (2021)
html
The strengths and weaknesses of DeFi price oracles. HackerNoon (2021)
html
Understanding the zero-knowledge proof method. HackerNoon (2021)
html
Most effective methods of optimizing gas when minting NFTs on Ethereum. HackerNoon (2022)
html

All essays and technical notes →

Experience

Since 2015Co-founder, Platinum Software Development Company / Platinum VC & Incubator (APAC) — software development, Web3 infrastructure, fund and incubator
2017 – 2018Chief Engineer, Lending & Risk Management, Everex (Singapore) — blockchain credit scoring, stable-value payments
2006 – 2015Sales & CRM-ERP Director, Merlion — one of the largest private IT distributors in Eastern Europe
2021Ph.D. in Education (Information Technologies: AI · zero-knowledge proofs · blockchain · data science)
2015M.Sc. in Information Security — Data Security and Cryptography, National Research Nuclear University MEPhI (Obninsk Institute of Atomic Energy)

Contact

Emaila@platinum.fund
Telegram@tonydzi
WhatsApp+1 341 222 9178
GitHubPalo-Alto-AI-Research-Lab
X / Twitter@antondzyatkovsk
LinkedInanton-dzyatkovskiy
ORCID0000-0001-7408-3054
Updated July 2026 · hand-written HTML, no framework, no tracking.