I build autonomous systems that operate, evaluate, and improve themselves without human micromanagement. Most recently: agentic AI infrastructure for a perpetual trading exchange, and before that, security automation for DeFi protocols deployed to mainnet.
Core work
- AI Agent Infrastructure — Multi-step orchestration, retrieval memory, tool-calling via MCP, structured generation (Zod), self-evaluation harnesses running 24/7 on Railway.
- Smart Contract Security — Static analysis automation (Slither, Foundry), adversarial pattern matching against 500+ historical exploits, autonomous audit pipelines.
- Backend Engineering — Rust, TypeScript, Python. APIs, settlement flows, infra-as-code (AWS, Terraform).
Stack
- AI / LLM: Anthropic SDK, OpenAI SDK, MCP, tool calling, vector databases, RAG pipelines, LangChain, LangGraph, multi-agent orchestration, context engineering, structured outputs, eval frameworks.
- Backend: Python, TypeScript, Node.js, AWS, GCP, Railway, Postgres, MongoDB, SQLite, Docker, REST / GraphQL APIs
- Blockchain: Solidity, Rust (Anchor), Move, ERC-4337/3643/4626/8004/8183, Foundry, Hardhat, smart-contract security