Computer Science student at the University of Minnesota Twin Cities
Building reliable ML systems, computer-vision experiments, and production-minded software
I am a Computer Science student interested in the engineering around machine learning: reproducible evaluation, data quality, failure handling, latency, memory, and interfaces that keep humans in control.
- B.S. Computer Science, University of Minnesota Twin Cities · Expected May 2028 · GPA 3.93
- Software Engineering Intern at FPT Software
- Contributor to The Shape of Noise, accepted to CTB at ICML 2026
| Project | What it demonstrates | Status |
|---|---|---|
| NoteFlow AI | FastAPI and React workflow for reviewable ASR/OCR records, comparisons, audit history, tasks, and exports | Active prototype; synthetic data only |
| Aerial Object Detection Benchmark | Reproducible VisDrone evaluation scaffold for CNN, DETR, Vision Mamba, and RT-DETR families | CPU-validated framework; GPU benchmark runs and results pending |
| ML Engineering Portfolio | Accessible, tested portfolio with verified case studies and automated GitHub Pages deployment | Live site |
Layer-wise perturbation profiles for diagnosing vision robustness. My contributions include controlled ResNet-50 and ConvNeXt-Tiny experiments, LoRA comparisons, layer-subset studies, and multi-seed analysis.
Topics: computer vision, robustness, PyTorch, LoRA, reproducible experimentation
- Efficient and hardware-aware ML systems
- Computer vision and robustness evaluation
- Reproducible experiments and benchmark design
- Backend and full-stack engineering for ML products
- Human-review workflows for model-assisted systems
Languages: Python, C++, Java, TypeScript
ML and vision: PyTorch, scikit-learn, OpenCV
Applications: FastAPI, React, SQLAlchemy
Workflow: Git, GitHub Actions, testing, experiment tracking, LaTeX
The best overview of my work is my portfolio. For academic or internship-related conversations, email me at quangminhph07@gmail.com.