Cornell Data Science

CDS Compute Cluster

This project designs, implements, and rigorously evaluates a learned cache eviction policy within a custom-built cache system, benchmarking its performance against traditional heuristics across diverse workload patterns.

LAN-first peer-to-peer distributed downloader for game/software installs and updates

World Wide Web 2 teaches an AI to browse the web by training it inside a learned simulator instead of a real browser. This makes training much faster and lets the agent learn from far more practice without needing live websites.

Python ONNX implementation of quantization techniques on machine learning models deployed to a Jetson Nano edge device.