AI engineer working mostly in Python. I build LLM-powered systems: agents, RAG pipelines, evaluation infrastructure, and the automation around them.
Working professionally since 2021, currently as a full-stack automation engineer, with freelance projects on the side.
RLHF and model evaluation. Authoring benchmark tasks and evaluation infrastructure used to train and test large language models, including agentic tasks with automated grading.
Production RAG agent. A hierarchical agentic RAG system that answers questions from real documents and structured data. Built for production, not a demo: retrieval routing, escalation to humans, and evals before anything ships.
API-driven automation. Backend automation running in production since 2021: third-party API integrations, webhooks, and internal tooling that replaced work people used to do by hand.
Algorithmic trading bot. A trend-following crypto bot (Donchian breakout, 4h timeframe) built on Freqtrade and deployed with Docker, plus a parallel news-watcher service that uses LLM classification to pause trading around macro events. Running in dry-run while the strategy proves itself.
RyxoGo. An experimental Go-to-WebAssembly frontend framework.
SecureChat. Real-time messaging app on the MERN stack with Socket.IO.
I like working in the open and contributing back. Right now I'm getting involved in open source LLM tooling, starting with LiteLLM. If you maintain something in that space and could use a hand, my inbox is open.
Python first. FastAPI and Node.js/TypeScript on the backend, React on the front, Go for side projects. MongoDB and Postgres for data, LLM APIs (Anthropic, OpenAI) for the agent work. I reach for plain Python before adding a framework layer that doesn't earn its keep.
