Agentic AI
Planning, tool use, and human oversight.
A closer look01Technology
From knowledge to working systems.
An interest in how AI systems act, find relevant information, and connect knowledge. The technical foundations matter as much as the experience of using them.
Core interests
RAG
Retrieval and generation grounded in relevant information.
A closer lookGraphRAG
Relationships, context, and knowledge graphs.
A closer lookSupporting disciplines
AI EngineeringData & AutomationWeb & AI Interfaces
↗Selected work
Ideas, put into practice.
Selected work spans LLM inference tooling, Vietnamese visual question answering, and computer vision experiments.
Open-source developer tool · 2026
InferCap
A toolkit for checking whether an LLM can run on available hardware, generating a vLLM serving configuration, and measuring inference capacity.
LLM Inference · Systems · ToolingOpen case studyCourse project · 2025
ViInfographicsVQA
A Vietnamese infographic question-answering project that follows the path from visual data to a working web demo.
Multimodal AI · VietnameseOpen case studyResearch exploration
Facial Landmark Localization
A computer vision study that turns a FaceXFormer survey into experiments, notebooks, and a documented evaluation path.
Computer Vision · FaceXFormerOpen case studyApplied AI systems project
Healthcare GraphRAG
A healthcare GraphRAG system combining semantic retrieval with structural search over a medical knowledge graph.
GraphRAG · Applied AI · HealthcareOpen case studyParallel programming final project
CUDA Autoencoder
A CIFAR-10 autoencoder with unified CPU and CUDA GPU pipelines, including naive GPU V1 and optimized GPU V2 kernels.
CUDA · GPU Systems · PerformanceOpen case study