Nguyễn Đôn
Quốc Tuấn.

Featured Work

AI Support Operations leads my work in full-stack AI: knowledge retrieval, cited answers, and human review. More projects follow below.

AI Support Operations

Personal project

A multi-tenant support platform that drafts answers from an approved knowledge base, with citations and human escalation.

Workflow

1. Knowledge base
2. Retrieval
3. Cited draft
4. Human review

• Implemented document ingestion, embeddings, and RAG retrieval with PostgreSQL/pgvector.
• Generated response drafts with source citations for human review.

More implementation details

• Added ticket classification, prioritization, and human escalation.
• Implemented workspace-scoped access control and team management.

React · TypeScript · FastAPI · PostgreSQL · pgvector · Gemini · Docker

Source code

IncidentLens

Full-stack application · Not yet runtime-validated

AI-assisted driving incident review using event logs, timelines, and video evidence.

Workflow

1. Video
2. AI observations
3. Evidence
4. Reviewer approval

• Integrated Gemini video analysis to generate timestamped observations linked to supporting evidence.
• Implemented engineer/reviewer roles and approval workflows for AI-generated findings.

More implementation details

• Integrated private video storage with temporary signed URLs.

Implementation prototype. The application has not yet been run or validated. AI observations require human approval; model accuracy has not been established.

Python · FastAPI · React · TypeScript · Supabase · Gemini · OpenCV · Docker

Source code

Build Phase Model - Assisted QC

AI Thực Chiến · Vingroup & VinUniversity · Submitted project

Model-assisted annotation QC integrated into CVAT, flagging missed objects and incorrect class labels for review.

Workflow

1. Annotations
2. Model checks
3. Human review

• Integrated model-assisted checks directly into the CVAT annotation workflow.
• Flagged missed objects and incorrect class labels for human review and correction.

CVAT · Computer Vision · Object Detection · Annotation QC

More projects

Student Attendance System

Developed an AI-powered student attendance management system using facial recognition technology. The platform enables lecturers to manage classes, record attendance automatically, generate reports, and monitor student attendance in real time through a modern web interface.

Node.js, Express.js, MongoDB, JWT, Swagger, Docker, Python, Flask, Face Recognition

Source code

Vietnamese Traffic Sign Recognition

Built a computer vision system capable of detecting and classifying Vietnamese traffic signs using YOLOv8. The system performs real-time inference on images and videos to support intelligent driver assistance and improve road safety.

Python, YOLOv8, OpenCV, Deep Learning

Source code

GPS Map Camera

Developed a GPS Map Camera application that captures images together with location, weather, compass direction, address, altitude, and timestamp information. The project includes a mobile application and cloud backend for media management.

React Native, Node.js, MongoDB, Supabase

Movie Recommendation System

Designed an intelligent movie recommendation platform using content-based filtering techniques with TF-IDF and cosine similarity, providing personalized movie suggestions.

Flutter, Flask, Python, Machine Learning

Source code

Experience

Skills & Expertise

Technologies I have worked with. The projects above show how I apply them.

Awards & Certificates

University recognition for my traffic sign recognition research.

Dong A University

3rd Prize - Scientific Research

Awarded for the AI Computer Vision Traffic Sign Recognition System.

View certificate

Dong A University

Scientific Research Recognition

Recognition for the traffic sign recognition research project in the 2025-2026 academic year.

View certificate

Services

The areas I work in, from AI models to the interfaces and APIs that make them useful.

AI Development

Building intelligent solutions using machine learning, deep learning, and computer vision technologies.

Web Development

Creating modern, responsive web applications with React, Next.js, and cutting-edge frontend technologies.

Backend Development

Designing scalable server-side architectures with Node.js, Spring Boot, and RESTful APIs.