What I Build
How I Work
I work with founders and small teams to build production-grade AI systems — from the first prototype to a robust, deployable product. I care about correctness, legibility, and craft: systems that do what they say, fail gracefully (no silent breakages, no mystery errors), and are easy to reason about. I bring both the depth to architect complex pipelines and the pragmatism to ship things that work.
Recent Projects
Consultant developer for core MVP of a regulatory intelligence platform. Built LLM-based horizon scanning for regulatory change detection, document analysis pipeline (PDF parsing, embeddings, semantic search), AI chatbot for policy Q&A, and compliance gap analysis.
Python · FastAPI · PostgreSQL/pgvector · Claude · Vertex AI · GCP · Firebase
Built Sitespry — a RAG-powered document intelligence product. Ingests PDFs, Word, websites; generates semantic embeddings and auto Q&A pairs; provides hybrid vector search with citations; and deploys as embeddable chat widgets. Plugin architecture, MCP server integration.
Python · FastAPI · Cohere · Gemini · libSQL/Turso · Cloudflare · Docker
Geospatial analysis (OS OpenData), social network analysis, vector embeddings (HyDE) in browser-based ML, Firebase auth on Cloudflare Workers. Published as technical articles.
Python · JavaScript · Firebase · Cloudflare Workers · Transformers.js
Tools I Use
Python · FastAPI · PostgreSQL · Claude · Gemini · Cohere · Vector DBs · GCP · Cloudflare · Docker