AIinMinutes/Data-and-AI-Concepts

Organizes 50+ executable Jupyter notebooks spanning transformer architectures (causal attention, multi-head attention), matrix calculus fundamentals, variational autoencoders, and classical ML/statistics concepts—each notebook provides hands-on implementations rather than theory alone. Covers a progression from foundational mathematics (Chebyshev's inequality, Jacobians) through modern generative AI techniques (GELU, perplexity, decoding strategies) and interpretability methods (SHAP, counterfactuals). Content targets practitioners preparing for data science roles with practical code examples across scikit-learn, PyTorch, and statistical libraries.

166 stars. No commits in the last 6 months.