I build Python automation, AI-assisted tools, and production-ready data pipelines. I like systems that are fast, reliable, and easy to debug.
I use AI tools like Cursor and Continue to move faster, but I keep full control over architecture, edge cases, and final quality.
⁛ Selected projects
- Panopticon CVE Map — local desktop app for vulnerability analysis with an interactive graph and a local AI agent (MCP + OpenRouter).
- SanityBot — real-time monitoring and automation system with WebSocket primary flow, REST fallback, atomic writes, and rate-limit protection.
- GooEye — async data collection and LLM pipeline with structured output, translation, and Telegram observability.
- Market Analytics Bot — real-time arbitrage signals bot with smart filters and an MCP server exposing market data as agent tools.
⁛ What I enjoy building
- Async systems that survive edge cases (WebSocket + REST fallback, retry with backoff)
- AI agents with structured tools (MCP servers, function calling, evals)
- Data pipelines that don't lose data (atomic writes, deduplication, cleanup)
- Desktop apps with local-first storage (SQLite, no cloud dependency)
- Systems with contracts — clear inputs, predictable outputs, no magic
⁛ Currently working on
- 🔍 Panopticon CVE Map — small fixes for stable version
- 📚 Learning: LangGraph, advanced async patterns, system design, Rust