At Darwin AI, I lead development of Darwin Workflows, the in-house automation engine that replaced third-party platforms and is now the default for every new account. I build it in TypeScript/Node.js on AWS, shipping features such as step retries with replay, signed failure-event webhooks and fixed-egress HTTP steps.
Before becoming a Software Engineer, I joined Darwin AI as a Support & Integrations Engineer and was promoted to Lead. I designed complex automations and built 60+ across metrics pipelines, AI agents, bots and platform integrations, while the team served a base of 600+ clients.
For my university thesis, I authored Nebula, a self-hosted semantic LLM gateway built with FastAPI, PostgreSQL and Qdrant. It routed requests across local, economy and frontier models, and measured 31% lower cost than routing everything to GPT-4.1 at equal quality.

