## Role Overview As a freelance Full Stack AI Engineer, I architect and ship production-grade AI-powered web products for early-stage startups and technical teams. Engagements range from 4-week MVP builds to ongoing fractional CTO arrangements. ## Key Deliverables - **RAG Starter Kit** — built a production-ready Next.js 15 RAG platform with hybrid Reciprocal Rank Fusion retrieval, pgvector, SAML 2.0 SSO, Inngest background jobs, and full OpenTelemetry tracing. Deployed on Vercel with one-click setup. - **Frontend Dev Helper** — shipped a Chrome/Firefox MV3 browser extension consolidating 27 debugging tools with Shadow DOM isolation, Canvas flame graph, AI-powered DOM analysis, and Stripe billing. Published to Chrome Web Store. - **React Perf Profiler** — developed a DevTools extension hooking into `__REACT_DEVTOOLS_GLOBAL_HOOK__` for React fiber-level render analysis, with three Web Workers, TensorFlow.js prediction, and a CI performance budget CLI. ## Technical Approach Every engagement starts with architecture — selecting the right abstractions for the problem rather than defaulting to familiar patterns. On the AI side, this means choosing between RAG vs fine-tuning, vector vs keyword vs hybrid search, and streaming vs batch generation based on the actual user interaction model. On the infrastructure side, it means making deliberate decisions about caching layers, background processing, and observability from day one rather than bolting them on later.