Systems boot. Code follows.

I build AI products and production systems from specs, contracts, tests, deploy proof, and tight feedback loops. Language is detail. Engineering is not.

boot order
spec -> contracts -> tests -> deploy -> smoke -> feedback

DentAI, UNIVALI, and AI stack in one graph.

Current work sits where clinical AI, academic platforms, and agentic tooling share one concern: reliable systems under pressure.

DentAI AI-first dental platform: RAG, Gemini, clinical assistant workflows, automation, release discipline.
UNIVALI Academic systems with Next.js, TypeScript, internal platforms, and team-level engineering flow.
AI stack Gemini, Claude, Codex, retrieval, evaluation, contracts, and smoke checks as production machinery.

Interfaces small. Contracts explicit. Proof hard.

Seniority is deciding what cannot stay vague before code starts moving.

AI products Workflows, evaluation, data boundaries, and user outcomes before model spectacle.
Backend systems Clear ownership, observable failures, stable contracts, and persistence that matches user reality.
Frontend flows State, API behavior, and UX shaped together so users do not feel machinery.
Release discipline Tests, CI, deploy visibility, smoke checks, and root-cause fixes instead of hope.

If you care about systems, AI, or sharp feedback, DM is open.

active stack: TypeScript / Next.js / React / Node.js / C# / PostgreSQL / Docker / AWS / GitHub Actions / Gemini / Claude / Codex / RAG