I'm a computer engineering student at Penn focused on thoughtful, reliable software across the web and machine learning. At Penn Labs, I lead frontend work on products used by thousands of students; at the U.S. Naval Research Laboratory, I've built computer vision tools and offline multimodal AI pipelines for scientific research. Outside of code, I enjoy gardening, weightlifting, reading, hiking, and playing piano, and I care deeply about the right to repair.
Summer 2026
A fully offline multimodal AI pipeline that extracts condition-level experimental parameters from academic PDFs using a QLoRA-fine-tuned Llama 3.1 8B model. A deterministic calculation engine applies confidence gating and null propagation, reducing paper processing from 1–2 hours to 6–7 minutes while preserving human review.
May 2026 — Jun 2026
A semantic search engine indexing 6,200+ mechanical keyboard listings from seven vendors. A multithreaded Python pipeline uses Groq and Gemini to normalize unstructured vendor data, while Jina embeddings and PostgreSQL pgvector power category-aware retrieval with an automatic SQL keyword fallback.
Jun 2026 — Present
A real-time shared canvas for friend groups, synchronized with Firebase Firestore snapshot listeners. Members can collect notes, bookmarks, images, and PDFs in a responsive masonry grid secured by Google OAuth, email whitelisting, and Edge Middleware.
2024 — 2025
An image analysis platform combining Cellpose and OpenCV to extract ice-crystal morphology, cutting a manual workflow by more than 95%. A custom parser also recovers images and metadata from proprietary instrument files.
2025 — Present
An upcoming peer-to-peer marketplace for the Penn community. I built Zod-validated listing flows, reusable interface components, and sublet address search backed by Photon and OpenStreetMap.