ModelHawk — Static ML Notebook Audit Platform
Next.js prototype · Static ML audit engine · Gemini/Ollama AI examiner · LaTeX PDF reports
Personal AI/ML prototype that audits Jupyter notebooks through static analysis, detects project profile, data modality, machine-learning workflow signals, evaluation risks and portfolio-readiness indicators, then generates AI examiner notes through a Gemini API route with Ollama local fallback and deterministic fallback behaviour.
The full workflow currently runs best from a local development environment because PDF generation depends on LaTeX/pdflatex, and the local AI fallback depends on Ollama. A stable hosted demo is planned after improving the deployment strategy for PDF generation and local-service-dependent features.