Hoàng Đăng Nguyễn

I’m an AI Engineer focused on turning advanced AI capabilities into production-ready platforms. At CMC OpenAI, I lead the architecture and productization of an enterprise AI Platform for LLM/VLM access, agent orchestration, knowledge bases and RAG, tool integration, evaluation, observability, and governance.

I designed Doc-Agent, an agentic document processing solution for document classification, information extraction, schema normalization, cross-document validation, human review, and evidence-grounded outputs. I also translate customer and business workflows into deployable AI solution architectures, evaluation plans, epics, user stories, acceptance criteria, and prioritized engineering backlogs.

Before that, I built GPU-accelerated OCR, information extraction, and multimodal inference services using Triton Inference Server, vLLM, and gRPC. I led VLM/LLM adaptations for form parsing, table understanding, key–value extraction, and visual question answering, and I built end-to-end document processing pipelines using Django.

I care deeply about reliability and operational quality—improving deployment reliability and observability with GitLab CI/CD, Docker Swarm, Prometheus, and Grafana. I’m currently pursuing a Bachelor of Data Science and Artificial Intelligence (Advanced Program) at Hanoi University of Science and Technology, and I’ve earned recognition including first-place wins in Vietnamese AI and handwriting/text recognition challenges.