At QuoteKamili, I build an AI pipeline that turns emailed, PDF and handwritten quote requests into priced quotes. It achieved 98% field accuracy in internal evaluation, and each quote line links to its source for approval.
I architected QuoteKamili’s multi-tenant AWS platform and ship one React codebase across web, iOS, Android and desktop. Prices and totals are handled by tested code, not AI models.
At Sociale, I led engineering on a multi-tenant fintech platform for savings groups and credit unions, with LLM agents in production for board reporting, risk analysis and member support. I also integrated an AI credit-scoring engine into the lending workflow.
My work has included a no-code credential workflow platform at Blocktrust, a Python FastAPI backend at DirectEd Development Foundation, and open-source work on Notamperdata. I’ve also taught LLM application design, RAG and GraphRAG through the Kenya AI Challenge.

