This project analyzes a year of SLURM sacct job data from Stanford's Sherlock cluster—covering ~860,000 jobs—to produce anonymized JSON summaries of user activity, resource usage, job outcomes, and temporal patterns that power a public dashboard
This project lowers the cost and carbon footprint of GenAI by compressing prompts through a lightweight LLM and optional Chinese translation, reducing token usage by over 50% without sacrificing accuracy. Awarded 2nd place in the Sustainable Smart AI track at HackMIT 2025.
Helping Climactic VC develop their comprehensive Climate x AI Market map, which showcases over 200+ companies working at the intersection of artificial intelligence and climate solutions across various sectors including energy, construction, agriculture, and more.
Tackling the U.S. electricity grid interconnection queue. Developed a user-centric platform that visualizes geospatial data of generation stations, manages each developer's projects, and organizes said projects into clusters using proprietary algorithms.
An app that uses image recognition to identify whether an item should be composted, recycled, or belongs in the landfill, tackling the expensive and harmful issue of compost contamination.
Advanced search engine implementation using semantic analysis to improve course discovery and matching for Stanford University courses.