Selected work across real-time experimental systems, large-scale data analysis, computer vision, and LLM tools.
Featured · Real-time systems
A full-stack VR system processing 6-DOF motion tracking input at 360 FPS. It renders perspective-correct 3D scenes at 120 FPS with under 20 ms detection-to-projection latency, using custom shaders I wrote for real-time cube mapping and projection transformations.
Data analysis
The core of my PhD. I applied linear and non-linear modeling, correlational analyses, and dimensionality reduction to understand how the mouse visual cortex integrates vision and body motion, with an automated Snakemake pipeline handling terabytes of data from preprocessing through visualization.
Computer vision
A computer vision pipeline for 1-photon calcium imaging movies. It achieves accurate cell segmentation and fluorescence signal extraction with minimal user input, including demixing temporal activity from spatially overlapping cells and re-identifying cells across experimental sessions.
LLM & retrieval
A RAG system trained on open-source permaculture and sustainable development wikis, including Open Source Ecology and One Community Global. It aims to help individuals and ecovillages design regenerative, sustainable food systems, with an eventual scope including sustainable building and decentralized energy, communication, and water systems.
