I'm a Computer Science + Informatics student, Statistics minor, at UC Irvine. I like building things people actually use — whether that's contributing to an open-source scheduler my school runs on or shipping camera features to beta testers. Right now I'm focused on product engineering/management, and going deeper on AI/ML.
- 🔬 Applied ML Fellow, Chewy AI Studio — evaluating classifiers for AI-generated text detection under adversarial paraphrasing
- 📱 Product & Software Engineer, Framelight AI — real-time camera guidance and gesture control in React Native, shipped to 100+ beta testers
- 🤖 AI/ML Fellow, Break Through Tech — ai/ml pipelines and mentorship
- 🗓️ Software Contributor, AntAlmanac — UC Irvine's course scheduling tool
Boring Notch turns the MacBook notch into a useful widget bar, but the upstream backlog of issues and PRs had stalled. Rather than wait, I forked it and started shipping the fixes and features myself.
- Real-time audio visualizer — plus a smoother, more accurate scrubbing bar for the music progress
- Clipboard section — expandable, with per-item delete
- Timer & stopwatch — built directly into the notch
- Camera over calendar — swapped the default calendar view for a live camera panel, moving calendar elsewhere
- Sharing/storage panel — horizontal scroll with per-item delete ("x")
-
Mindtrail Tool-calling research agent with vector-embedded memory; 70% recall@1 on a 130+ test evaluation harness, containerized with Docker.
-
FraudGuard Fraud detection platform with XGBoost, FastAPI, React, and SHAP explainability.
-
hospital-readmission-prediction Ensemble models predicting 30-day readmissions from EHR data, with SHAP explainability.
-
AntAlmanac Course exploration and scheduling tool used by UC Irvine students — active contributor, work in progress.
-
ml-flashpoint — Google's open-source ML checkpointing library Refactored the Megatron adapter, extracting save logic into a reusable helper (66% fewer lines), and abstracted 5
torch.distributedAPIs via dependency injection to enable deterministic unit testing across 5 test files.
Python | TypeScript | JavaScript | Java | C/C++ | SQL | R
PyTorch | Scikit-learn | LLMs | RAG (LlamaIndex, FAISS) | SHAP | MLflow | OpenCV
React | React Native | FastAPI | Node.js | Docker | AWS | Supabase | Firebase | Git