Volume 1 · 20 chapters
Classical Data Science
The foundations. Framing the business problem, the maths, Python and SQL. Then the classical ML toolkit: ensembles, validation, features, evaluation and time series.
Series
A field guide for solutions architects moving into data science and AI. Four volumes, 83 chapters, from statistics to world models. Every chapter pairs the ideas with trade-offs, a worked example and code you can run.
Volume 1 · 20 chapters
The foundations. Framing the business problem, the maths, Python and SQL. Then the classical ML toolkit: ensembles, validation, features, evaluation and time series.
Volume 2 · 17 chapters
From MLOps to LLMs. Responsible AI, deployment and monitoring. Then neural networks, transformers, RAG, fine-tuning, RLHF and agents.
Volume 3 · 20 chapters
Making it fast. GPU architecture, CUDA and Triton kernels. Then CPU performance: profiling, SIMD, parallelism and memory. Ends with an end-to-end optimization.
Volume 4 · 26 chapters
Three frontiers. Physical AI: perception, world models, VLAs and robot policies. Image models, from CNNs to diffusion. Knowledge graphs, GraphRAG and agentic reasoning.
The books are licensed under CC BY-SA 4.0. You may copy, share and adapt them, including for commercial use. Credit the author, and share any changed version under the same license. The companion notebooks are licensed under the MIT License.