Available for opportunities · 2025

Son Hong
Nguyen

Computer Vision Engineer · AI Engineer

B.Sc. Data Science · HCMIU · Vietnam

Specializing in computer vision — image classification, object detection, segmentation, and style transfer. Strong background in model development and production deployment. Published researcher with 2 submitted journal papers in CV and deep learning.

nguyenhongson0305@outlook.com +84 949 682 766 LinkedIn GitHub

Download CV View Projects

AI · CV

Technical Profile

Skills & Toolset

Programming

Python SQL

ML / Deep Learning

PyTorch OpenCV NumPy scikit-learn

Computer Vision

Classification Detection Segmentation Style Transfer Face Recognition

CV Models & Libs

MTCNN FaceNet MediaPipe ResNet AdaIN

Backend & Deployment

Flask FastAPI SQLite PostgreSQL Gunicorn Nginx Linux VPS

Tools & Infra

Git Linux NVIDIA Jetson Nano SMB Protocol

Work History

Experience

Jul 2024 – Oct 2024

Velox Tech Company

AI Engineer Intern

Cleaned, preprocessed, and curated datasets for model training pipelines.
Analyzed model outputs and error modes; recommended architecture and data improvements.
Designed and implemented a distributed PostgreSQL database to synchronize data between lab and environment systems.

Academic Background

Education

Bachelor of Science in Data Science

HCMIU — Vietnam National University Ho Chi Minh City

2021 – 2025

Relevant coursework: Machine Learning · Deep Learning · Computer Vision · Statistics · Data Mining · Data Analysis

Research

Publications & Papers

ORCID 0009-0009-3767-124X

Submitted · 1st Author

Learned Weighted Hypotheses Union for Robust Prompt-based Segmentation

Submitted to Engineering Applications of Artificial Intelligence — Elsevier · IFAC

Submitted · 1st Author

Enhancing Multi-Region Stylization with Interior-Guided Boundary Repair

Submitted to Multimedia Systems — Springer Nature

Work

Featured Projects

📄 Research

🧠

Computer Vision · Deep Learning

Face Recognition System v3 — FRS

Biometric authentication with liveness detection (Silent-Face CNN ensemble), iris gaze tracking (MediaPipe), and 512-d FaceNet embedding. MTCNN face detection with cosine similarity matching.

PyTorchMTCNNFaceNet MediaPipeFastAPIOpenCV

View Details

📄 2 Papers

🎨

Computer Vision · Style Transfer · NLP

Prompt-guided Multi-Region Image Style Transfer

Segments objects from text prompts and applies AdaIN style transfer per region. Includes interior-guided boundary repair for seamless multi-region stylization. Subject of 2 submitted journal papers.

PyTorchAdaINSAM CLIPOpenCV

Research Project

● Live

💾

Full-Stack · Computer Vision · NAS

SonDrive — Personal Cloud File Manager

Self-hosted web file manager with AI-powered biometric login — face recognition + liveness detection + gaze tracking. SMB NAS backend. Media streaming, ZIP download, shareable links, PDF preview.

FastAPIFaceNetsmbprotocol SQLiteWebRTC

View Details ● Live

📦

Full-Stack · Data Engineering

Warehouse Management System — WMS v1.0

Multi-tenant inventory platform with per-user isolated SQLite databases. Real-time stock tracking, time-series analytics (week/month/year), full audit trail with JSON diffs, employee scheduling.

PythonFlaskSQLite PBKDF2REST API

View Details ● Live

🎯

Full-Stack · Productivity · Real-time

TaskMaster v2 — Deep Work OS

Productivity platform with Eisenhower matrix, Pomodoro deep-work sessions, streak tracking, group collaboration with SSE real-time push, democratic task voting, and behavioral analytics schema.

PythonFlaskSSE SQLiteArgon2

View Details

🤖

Edge AI · Embedded Systems

Real-time Human Alert System

Deployed an optimized detection model on NVIDIA Jetson Nano to monitor specific regions and trigger real-time alerts. Model optimization for edge inference with limited compute budget.

PyTorchOpenCV Jetson NanoNVIDIA

Edge Deployment

🏗️

ResNet from Scratch

Implemented ResNet architectures from scratch in PyTorch and applied to image classification benchmarks.

Son Hong Nguyen