Hoang Nguyen | AI & Software Engineer

About Me

Engineer who ships AI that cuts through the noise.

I close the gap between research and production — building AI systems that are reliable, scalable and actually used.

📌 Background

Who I Am

AI engineer with an M.S. from La Trobe University and a B.Eng. in Computer Engineering from HCMUT. I've built multi-agent pipelines, RAG systems, and cloud-deployed APIs — and published first-author research in Human Factors (IF 3.3). I care about clean architecture, reproducibility, and writing code that holds up in production. Off the clock, you'll find me out running — the perfect pace for thinking through hard problems.

🏃 Follow on Strava

✦ What I Bring

My Edge

I bring a unique combination of deep AI/ML expertise and practical engineering skills. From designing multi-agent pipelines and RAG systems to deploying robust cloud APIs, I ensure that every solution is not only innovative but also production-ready.

"I don't just build models — I build systems that work in production and create measurable impact for the teams I join."

— HOANG NGUYEN

Experience

Working Experience

AI Research Assistant

La Trobe University

📍 Melbourne, Australia

2024 – 2025

Collaborated with Dr. Huawei Tu at La Trobe University on AI/ML and deep learning research for object recognition using hardware sensor data, applied to an immersive VR headset experience. Conducted experiments, analysed motion sensor datasets, and co-authored a peer-reviewed paper published in Human Factors (Impact Factor 3.3) as part of the M.S. in AI program.

📜 Published Paper — Human Factors (IF 3.3)

Software Developer

Sun.Studio

📍 Ho Chi Minh City, Vietnam

2022 – 2024

Developed and shipped top-grossing games on Facebook and multiple platforms. Built and maintained game frontend UI, collaborating with designers and artists on interface design. Applied AI-driven procedural generation and math/physics logic optimisation to improve performance and deliver unlimited maps.

🎮 Bubble Fall 🎮 Solitaire 🎮 Solitaire Tripeaks Journey

Projects

Proficient Projects

⭐ Top Project

🤖

LLM RAG LangChain FAISS HuggingFace FastAPI PostgreSQL SQLAlchemy

Multi-Agent AI Assistant

Fully offline-capable multi-agent AI assistant using a hierarchical CodeAgent architecture (smolagents/HuggingFace). A Manager Agent orchestrates a Web Search Agent (DuckDuckGo) and a RAG-based Retriever Agent, delegating queries intelligently across both. LLM inference runs locally via Ollama. RAG pipeline uses HuggingFace tokenizer-aware chunking, thenlper/gte-small embeddings, and a persisted FAISS vector index. Chat history stored in PostgreSQL via async SQLAlchemy; responses stream via SSE. Fully Dockerized with docker-compose.

▶ View on GitHub ★ … stars

🥼

🎓 Master Thesis · Score: 89 / 100

PyTorch Scikit-learn Unity VR Decision Tree ExtraTrees

Arm Tapping Recognition with Two-Handed VR Controllers

Developed a real-time VR arm-tap recognition system achieving 95% classification accuracy using PyTorch and Scikit-learn. Engineered motion-based features from controller sensor signals and trained Decision Tree and ExtraTrees models with hyperparameter tuning and cross-validation. Deployed the optimised model into Unity for low-latency inference, enabling reliable gesture interaction without additional hardware sensors.

🎯 95% accuracy 📄 View Thesis

📝

Python OpenAI API FastAPI

AI Job Application Assistant

LLM-powered automation system that analyzes resumes against job descriptions and generates structured fit scores and tailored cover letters. Designed prompt templates with JSON schema validation for reliable outputs and implemented a modular pipeline for scalable evaluation workflows.

▶ View on GitHub

🧾

FastAPI SQLite AWS EC2/S3 Docker Nginx

OCR Standardized Invoice Service

Cloud-based REST API to convert OCR outputs into standardized invoice JSON schemas. Built customizable field-mapping and validation layers to ensure data consistency across varying invoice formats. Implemented SQLite storage, structured logging, health checks, and deployed using Docker, Nginx, and AWS EC2 with S3 integration.

▶ View on GitHub

🔬

PyTorch U-Net Albumentations

Image Segmentation with U-Net

Medical image segmentation system using U-Net (31M parameters) in PyTorch. Built a complete ML pipeline including custom dataset handling, data augmentation (Albumentations), multiple loss functions (BCE, Dice), and automated training with validation monitoring.

▶ View on GitHub

Skills

Technical Expertise

🧠

AI & ML

Building and deploying end-to-end AI systems — from LLM fine-tuning and RAG pipelines to deep learning and computer vision models.

LLM RAG FAISS NLP Deep Learning Hugging Face

🛠️

Frameworks & Tools

End-to-end tooling for AI backends and agent systems — API development, orchestration, containerisation and retrieval.

FastAPI SQLAlchemy LangChain Docker smolagents

💻

Languages

Proficient across systems, scripting and application languages — from AI backends to embedded and full-stack development.

Python C++ C# C TypeScript / JS

☁️

Cloud & DevOps

Deploying and managing production AI services on cloud infrastructure with CI/CD, containerisation and reverse proxying.

CI/CD Nginx AWS Docker

🗄️

Databases

Designing and querying relational and managed databases for AI backends — schema design, query optimisation and cloud storage.

PostgreSQL SQLite Supabase

Hello, I'm
Hoang Nguyen

Engineer who ships AI that cuts through the noise.

Who I Am

My Edge

Academic Background

Working Experience

Proficient Projects

Technical Expertise

Language Proficiency

Let's build something
great together.

Hello, I'mHoang Nguyen

Engineer who ships AI that cuts through the noise.

Who I Am

My Edge

Academic Background

Working Experience

Proficient Projects

Technical Expertise

Language Proficiency

Let's build somethinggreat together.

Hello, I'm
Hoang Nguyen

Let's build something
great together.