Nguyen Dang Trac
AI Automation Engineer
I build AI-native products end to end — multi-agent systems, RAG pipelines, and multi-tenant platforms that ship to production.
What I do
M.S. in Electrical Engineering with three years spanning AI research, automotive embedded software at Bosch, and production-grade AI-native product engineering at Edge8 AI. I build from-scratch multi-agent systems, retrieval pipelines, streaming AI UX, and clinical risk engines — the whole delivery lifecycle, from architecture to test coverage.
AI & Agentic Systems
From-scratch multi-agent orchestration (router → planner → executor), custom multi-model LLM clients, streaming reasoning, and MCP tool layers — no framework crutches when they get in the way.
Retrieval & Search
End-to-end RAG: multi-strategy chunking, pgvector + FAISS + BM25 hybrid retrieval, and benchmark-driven routing decisions measured on production corpora.
Full-Stack Product
Next.js App Router frontends and FastAPI backends on Supabase and Vercel, with multi-tenant isolation, real-time streaming UX, i18n, billing, and E2E test suites.
Selected work
Research
Three IEEE publications in deep learning for error-correction coding, including a first-author IEEE Transactions journal and a Best Paper Award.
D.-T. Nguyen & S. Kim, “U-shaped Error Correction Code Transformers” ↗
T. A. Khoa et al., “Safety Is Our Friend: A Federated Learning Framework Toward Driver's State and Behavior Detection”