Fullstack AI Engineer building production-ready AI applications end to end — from polished interfaces and backend infrastructure to evaluation, deployment, and operations.
I work on systems where AI has to hold up in real workflows: evidence-led research, document AI, agent tooling with explicit control boundaries, human-in-the-loop computer vision, and optimization software.
Currently building AI applications for technical documentation at Kothes.
Abundance — evidence-led deep research with explicit counterevidence, calibrated confidence, traceable sources, and a deterministic LangGraph workflow.
Optical Context MCP — compresses OCR-heavy PDFs into dense packed images for vision-capable agents. A documented validation run reduced 22 source pages to 6 retrievable images. Available on PyPI.
Sewer Inspection AI — AI-assisted defect detection for sewer CCTV footage, combining YOLO, overlay OCR, finding aggregation, and a review cockpit for human sign-off.
Conversational Production Optimizer — connects a natural-language planning interface to deterministic GAMS optimization, turning operational constraints into mathematically optimized production plans.
KLICKin — career platform with 550+ registered users, real-time chat, OAuth, admin 2FA, and automated deployments with health checks and rollback. Private repository; architecture walkthrough available on request.
Python · FastAPI · LangChain · LangGraph · Next.js · React · TypeScript · PostgreSQL · Docker