Coursework included data structures, algorithms, computer architecture, operating systems, computer networks, and software engineering.
First job out of university. Added networking to Quixel Bridge so one machine could act as a Megascans asset server for the rest, reducing redundant downloads across the office. Quixel Bridge is now part of Unreal Engine.
Stack: React, Node.js, Electron.
Two years of contract work across the stack: a Prismic + Gatsby blog, a React Native app for ListenField (Japan), and Dubseek, a music-streaming service for a US client.
Stack: React, Next.js, GraphQL, React Native, Gatsby, Postgres, Nest.js.
Two and a half years at a VC-funded iPaaS startup ($3M+ raised), spanning three roles:
Stack: React, Redux, TypeScript, Node.js, Express, Docker, Kubernetes, Mongo, SQL.
Built expertise in generative AI:
Leading the AI/ML layer of FanVoice, a voice-and-video feedback platform, driving it end to end across the backend, the media pipeline, and the capture and review apps.
Took FanVoice's feedback chat from slow and expensive to instant and cheap: RAG over vector embeddings cut OpenAI costs ~90%.
FanVoice's flagship analysis feature: it turns a pile of raw feedback into structured, cited findings.
Resolve a user's issue before a support ticket is ever created. While they record or type, the system transcribes in real time and synthesizes a direct answer.
Built Elliot, the team's in-house AI engineering assistant. A Slack bot on the Claude Agent SDK that answers any question about the company's products and creates PRs, backed by a custom MCP server and a context-distribution CLI that keeps every engineer and AI tool working from the same, current knowledge.
A customer needed analysis in seconds, not minutes. A composite fast path, benchmarked against the old pipeline on real videos: 3× faster at the median, 4.6× at the tail.