Applied / Account Engineer · AI-native

Models, governed.
Shipped, not slideware.

I build LLM systems wrapped in deterministic, verifiable workflows — multi-model routing, non-LLM scoring, dataset-backed evals, human-in-the-loop, fail-closed safety. 55 projects · 38 AI-powered · 25 live. Every tile is a real preview or interactive demo. No client data, ever.

Claude · OpenAI / Codex · Groq · Gemini · local Ollama · Next.js · Supabase · Playwright · deterministic pipelines · dataset evals · fail-closed safety Claude · OpenAI / Codex · Groq · Gemini · local Ollama · Next.js · Supabase · Playwright · deterministic pipelines · dataset evals · fail-closed safety
Explore the work ↓ Résumé · 1-page Privacy & security audit
scroll
0
Buildable projects
0
AI-powered
0
Shipped & live
0
Models orchestrated

Selected work

hover a tile — it comes alive · click to open
All AI Pipelines Live Products Dashboards & Tools Agentic / Safety