nested-learning

Implement AI/ML research papers from scratch when no official code exists. Use when the user wants to reproduce a paper, implement an algorithm from a PDF, build a model architecture from a research paper, or create working code from academic publications. Handles papers from arXiv, NeurIPS, ICML, ICLR, CVPR, and other venues. Produces UV-managed, GPU-ready Python projects with tests, demos, and documentation.