I build AI products end to end, the LLM systems and data pipelines, the app around them, and the infrastructure that keeps it all running in production.

What I care about most is the unglamorous half of AI engineering: making generation systems reliable instead of just demo-ready, grounding them in real data, and keeping the infra and token cost under control once real users show up.

📍 France · Remote · 🌐 nicolaedogotaru.com · ✉️ nicolaedogotaru@gmail.com

• Design LLM-driven systems: generation, enrichment, retrieval —> built to hold up in production, not just in a demo.
• Build the backend and data flows around them: Next.js, Laravel, Node, Python.
• Run the infrastructure myself (k3s, Terraform, Docker, GitHub Actions). I'd rather own the whole chain than hand off the part that breaks first.

Languages — TypeScript, JavaScript, Python, PHP Front / Back — Next.js, React, Node, NestJS, Laravel, Django · PostgreSQL, Redis, Supabase, Prisma Infra & ops — Kubernetes / k3s, Docker, Terraform, GitHub Actions, Traefik, Cloudflare Monitoring — Grafana, Prometheus, Sentry, Splunk

STHO — an AI-native platform for the live-entertainment industry. I built the LLM generation systems at its core (automated venue programming, lead generation pairing scraping + LLM enrichment, reputation analysis), the product around them (Next.js + Laravel), and ran the whole cloud-native infrastructure.

EarnYourStream — a real-time bridge between TikTok Live and game servers: viewer actions trigger in-game events with no perceptible delay, on a Kubernetes setup built to absorb the spiky load of live streaming.

If you're building something where the AI has to actually work in production — not just look good in a demo — that's the kind of problem I like.

✉️ nicolaedogotaru@gmail.com · 🌐 nicolaedogotaru.com · github.com/ndogota