Hung Manh Hoang (Henry Hoang)
Fullstack AI Engineer at FPT Software
Hanoi, Vietnam · AWS Certified AI Practitioner
About
I'm an engineer specialising in AI agent and LLM application development. I've designed and delivered 5+ production AI/LLM systems across education, finance, technology and e-commerce — covering the full path from architecture to operation: retrieval-augmented generation, multi-agent orchestration, and the evaluation and monitoring that keep both dependable once real traffic arrives.
I architect and run these workloads on AWS and Azure, with a focus on systems that stay reliable, secure and scalable under load. What interests me most is the gap between a working demo and a production system, and the engineering discipline it takes to close it.
Currently finishing a Computer Science degree at Hanoi University of Industry. Open to conversations about retrieval quality, LLM evaluation, and practical multi-agent design.
What I work on
RAG systems
Retrieval pipelines end to end: document ingestion and chunking, embedding and vector indexing, hybrid retrieval with reranking, and grounded generation with source attribution. Retrieval quality measured against eval sets rather than assumed from demos.
Multi-agent systems
Specialised agents handling sub-tasks under an orchestration layer — tool use, state and context passing between agents, retry and fallback behaviour, and guardrails on what each agent is permitted to do.
Cloud and platform
Running these workloads on AWS and Azure: deployment and scaling, cost and latency management, access control and data handling, and tracing on every LLM call so production behaviour is observable rather than inferred.
Software engineering
Shipping the system around the model — services and APIs in Python, prompt and configuration versioning, and evaluation harnesses so changes are compared instead of guessed at.
Stack
- Python
- LLM APIs & open-weight models
- RAG
- Vector databases
- Agent frameworks
- AWS
- Azure
Find me elsewhere: