Computer Science @ UC San Diego · Break Through Tech AI/ML Fellow
I'm a computer science student at UC San Diego who enjoys building dependable software around machine learning and real-world data. I'm pursuing machine learning engineering internships, especially opportunities where I can evaluate models, build reliable data pipelines, and turn predictions into useful products.
Through Break Through Tech, team projects, and open-source work, I've gained experience carrying ideas from data preparation and experimentation through backend integration and interactive interfaces.
- Deepening my skills in model calibration, evaluation, and production ML pipelines
- Contributing to open-source backend and cloud-storage infrastructure
An end-to-end system for forecasting harmful algal bloom risk along the California coast.
- Processed 50,670 NOAA satellite observations and engineered rolling, anomaly, and lagged environmental features
- Trained and calibrated a Random Forest model on 3,078 weekly records, reaching approximately 0.86 ROC-AUC
- Built an automated weekly data pipeline and a Supabase-backed interactive risk dashboard
- Stack: Python, scikit-learn, XGBoost, Pandas, Supabase, GitHub Actions
- Live dashboard
A privacy-preserving occupancy system that performs inference from an 8×8 thermal sensor instead of a camera.
- Quantized a TensorFlow Lite Micro model to 6.5 KB and deployed it to an ESP32 with 90.2% accuracy
- Built the MQTT, FastAPI, MySQL, OAuth, and WebSocket pipeline behind a live monitoring interface
- Stack: C++, Python, TensorFlow Lite Micro, FastAPI, Docker
- Watch the demo
A multi-agent travel planner that coordinates flight, hotel, activity, and review workflows.
- Passed 20/20 live benchmark prompts and generated ranked itineraries in 15.5 seconds on average
- Won the SanD Hacks AGNTCY Track
- Stack: Python, LangGraph, OpenAI API, React, Docker
- Watch the demo
A hardware-software prototype that helps identify counterfeit sneakers using camera, spectral, weight, and shape signals.
- Engineered a two-Raspberry-Pi pipeline that combined camera, spectral, and load-cell data and returned an authentication result to an iOS app
- Helped earn 2nd place at the Art of Product Engineering Demo Day after measuring 12.3% weight and 5.2% spectral differences
- Stack: Python, FastAPI, Raspberry Pi, SwiftUI, WebSockets
- Watch the demo
Languages
Machine learning and data
Web, backend, and infrastructure
I contributed a complete Azure Blob Storage provider to Portabase, covering upload, download, delete, copy, connection testing, validation, React setup, provider registration, and a database migration.
- Break Through Tech AI/ML Fellow — selected from 4,000+ applicants
- SanD Hacks AGNTCY Track Winner
- ECE140B Demo Day — 2nd Place
- BASTA Code2Career Fellow
Away from my keyboard, I enjoy tennis and strength training. Both keep me patient with repetition and focused on steady improvement—the same mindset I bring to engineering.
- Email: j6min@ucsd.edu
- LinkedIn: linkedin.com/in/jingimin
- Portfolio: jingimin.dev