research

Research directions and selected work by Tran Phuoc Thanh Nhan in NLP, Generative AI, RAG, and Multi-Agent Systems.


My research focuses on building practical and reliable AI systems at the intersection of Natural Language Processing, Generative AI, Retrieval-Augmented Generation, and Multi-Agent Systems. I turn research ideas into usable systems, reproducible experiments, and project outcomes.

Research mission: Build AI systems that are grounded in evidence, useful in practice, and transparent enough to evaluate and improve.

Research snapshot

Focus areas Methods Outputs
NLP, Generative AI, RAG, and Multi-Agent Systems Retrieval, agent orchestration, LLM adaptation, and evaluation Research papers, prototypes, open-source projects, and technical writing

Research directions

Retrieval-Augmented Generation

I study how retrieval can improve the factuality, controllability, and domain adaptation of large language models.

What I study:

  • Hybrid search and document grounding
  • Query understanding and retrieval quality
  • Evaluation for academic and institutional question-answering systems

Related work: AskUIT AI Academic Assistant, RAG survey notes, and academic support chatbots.

Multi-Agent AI Systems

I explore how specialized agents can collaborate to solve information-seeking and reasoning tasks.

What I study:

  • Task decomposition and agent routing
  • Tool use, verification, and orchestration
  • Reliable workflows for domain-specific AI assistants

Related work: Multi-agent architectures for academic assistance and domain-specific AI workflows.

NLP and Generative AI Applications

I work on applied NLP systems that combine text understanding, generation, retrieval, and evaluation for real-world use cases.

What I study:

  • LLM fine-tuning and LoRA-based adaptation
  • Vietnamese and English NLP applications
  • Practical evaluation of generated content and AI system behavior

Related work: Publications and technical notes on RAG, LLMs, and AI system design.

Selected project outcomes

  • AskUIT AI Academic Assistant — a multi-agent RAG chatbot for academic support, recognized with an Encouragement Prize at SEApp 2025. See the Projects page for more technical work.
  • Competitive Programming Repository — a long-running collection of optimized algorithmic solutions and reusable C++ templates.
  • Research writing and surveys — technical writing on RAG, LLMs, and AI system design. Browse the Blog for research notes and reflections.

Explore more

  • Publications — papers, preprints, and citation metadata.
  • Projects — implementations, demos, repositories, and technical details.
  • Blog — research notes, tutorials, and technical reflections.