Building a tool that reviews AI-generated code for security issues and bad practices, explaining problems in plain language — for developers who use AI coding tools but don't fully understand every line it writes. Running on AMD GPUs. Looking for a teammate comfortable with Python/ML.