Vision Team, FPT SmartCloud (FCI), FPT

Leading the development of advanced document intelligence and data ingestion systems for healthcare and administrative documents. Building retrieval and classification systems for complex PDF processing.

Key Achievements:

• • Developed and deployed text-based retrieval and classification models for OCR Healthcare data (ICD-9, ICD-10, surgical datasets)
• • Designed and implemented advanced RAG-based system for question-answering on healthcare insurance PDF documents with module-level performance evaluation
• • Implemented modular, extensible Multimodal Data Ingestion Pipeline for administrative PDF/Docx/PPTX documents, delivering structured outputs in HTML and JSON formats
• • Developed and enhanced models addressing Language Detection, Table Merging, Text Mapping, Reading Order, and Document Structure Analysis challenges

Technologies:

Innovation Center, VNPT-IT, VNPT

Designed and built neural network models and retrieval systems for VNPT's SMARTBOT framework. Handled back-end engineering for scaling and improving performance of retrieval-based systems.

Key Achievements:

• • Designed neural networks models including intent classification, rule-based entity recognition, statistical entity recognition, out-of-vocabulary detection, and accent-corrector models for VNPT's SMARTBOT system
• • Designed legal document retrieval system and matching address algorithm for VNPT's eKYC integrated with OCR system
• • Built, served, scaled and improved performance of retrieval-based system using Python, Elasticsearch, FastAPI, AsyncIO, Redis and parallel programming techniques

Technologies:

Viettel Research and Development Institute

Worked on information extraction problems in natural language processing within a team environment.

Key Achievements:

• • Worked in a team of five people to solve Information Extraction problems in Natural Language Processing
• • Wrote technology internal paper on 'Supervised Methods for Named Entity Recognition problem'

HUST's Embedded System and Reconfigurable Lab - University of Aizu's Communication Lab

Research on machine learning approaches for video quality assessment.

Key Achievements:

• • Developed Machine learning models to solve Video Quality Assessment (VQA) problem about Video Streaming

Hanoi University of Science and Technology

Major: Electronics and Telecommunications - Talent Program

NLP & Language Models

Machine Learning

Document Intelligence

Backend & Infrastructure

Data Processing

Data Engineering

Cloud & Tools

VLSP 2025

Sub-Task 2 DurationQA - Applied retrieval and fine-tuning techniques to Qwen models

Trustii.io

Built, deployed offline, secure and efficient RAG System for Understand.Tech

Zindi / TechCabal

Trained on-edge CNN model for audio classification problem

Zalo AI

Designed Large Language Model (LLM) for Elementary Maths Solving

Zalo AI

Vietnamese Wiki Question Answering Challenge

Vietnam Olympiad of Informatics (Ministry of Education)

National informatics olympiad organized by Ministry of Education