George Perdrizet

Educational tool for Python bug-fixing. Students are shown AI-generated broken snippets and must find and fix the bug in-browser using a Monaco editor with Pyodide-powered test execution.

An authenticated, metered API gateway for LLM inference. Provides OpenAI-compatible endpoints with per-user rate limiting and usage tracking over a self-hosted llama.cpp backend.

A platform for delivering technical classes taught by me, with course materials, assignments, and progress tracking for data science and AI/ML topics.

Hide secret messages inside naturally-generated text using next-token probability steering. Encodes information into LLM outputs by biasing token selection at inference time.

A collection of containerized development environments for AI/ML and data science education, designed for use with VS Code Dev Containers. Provides reproducible, zero-config setups for students.

A Python package for hill climbing optimization of user-supplied objective functions, with simulated annealing and parallel replica exchange. Designed for flexible multi-objective optimization with multi-column dataset support.

An educational distributed deep learning system where students become part of a compute cluster to train a GAN. Teaches distributed ML concepts by making each student's machine a real training node.