ViInfographicsVQA
Team project covering Vietnamese infographic data, question–answer pairs, model exploration, and a Streamlit / FastAPI demonstration.
01 /AI SYSTEMS · SOFTWARE ENGINEERING
Working on AI systems, ML infrastructure, and production software.
Based in Ho Chi Minh City, Vietnam
I build useful AI and software from messy technical questions—and explain the decisions so people can trust and use the result.
02 /Selected work
Three bodies of work showing how I build AI systems, inference infrastructure, and production software.
LLM Inference · Systems · Tooling
A toolkit for checking whether an LLM can run on available hardware, generating a vLLM serving configuration, and measuring inference capacity.
CUDA · GPU Systems · Performance
A CIFAR-10 autoencoder with unified CPU and CUDA GPU pipelines, including naive GPU V1 and optimized GPU V2 kernels.
GraphRAG · Applied AI · Healthcare
A healthcare GraphRAG system combining semantic retrieval with structural search over a medical knowledge graph.
03 /TECHNICAL FOCUS
Technology is the main thread; other domains are context for the technical work.
Agentic AI, retrieval, and knowledge graphs for systems that can act and answer from knowledge.
Hardware checks, model serving configuration, benchmarking, and inference telemetry.
CLI workflows, endpoint verification, automation, and interfaces that make systems usable.
Multimodal question answering and computer vision experiments from data to inspectable demos.
04 /WRITING & RESEARCH
My academic projects include Vietnamese infographic question answering and facial landmark localization. They connect computer science with dataset preparation, model exploration, and evaluation.
Team project covering Vietnamese infographic data, question–answer pairs, model exploration, and a Streamlit / FastAPI demonstration.
Research exploration covering related work, FaceXFormer, smaller-dataset experiments, and evaluation notebooks.
05 /ABOUT THIEN NGUYEN
Nguyen Van Minh Thien
I work across AI, frontend systems, and computer science. I care about the whole trail of a project: framing the question, testing an approach, documenting the limits, and making the result easy to explore. Teaching programming and mathematics is a supporting interest, not the centre of this site.
View projects↗CONTACT
I’m open to project work, research discussion, and product-minded frontend collaboration. Reach out through GitHub, LinkedIn, or email.