A living reading on production AI.
Printed books on AI are obsolete before they ship. AI Daddy is the alternative: a living reading, written by an engineer shipping production AI in enterprise, that updates as the field moves.
What you'll find here.
Eighteen sections, more than a hundred chapters, covering foundations, model landscape, training, inference, prompting, retrieval, agents, memory, frameworks, document processing, infrastructure, security, reliability, evaluation, design patterns, case studies, and tool-use and computer agents.
Each chapter is opinionated about tradeoffs: latency versus cost, accuracy versus faithfulness, single-agent versus multi-agent. Where vendors and models matter, they are named, because real systems require real choices.
How it stays current.
The content lives in an open-source repository on GitHub. When a new model ships, a protocol changes, or a pattern emerges, the repo updates and the site rebuilds. Pull requests from practitioners are welcome. See the contributing guide.
Who writes it.
Maintained by Om Bharatiya, a Principal AI Engineer working on agentic AI platforms and evaluation infrastructure in enterprise banking.
How it's funded.
Every chapter is free to read online. The same chapters are also collected into paid PDF guides, sold on the guides page, and it's that revenue, plus a small number of unobtrusive ad slots, that keeps the lights on. If you'd like to sponsor a section, get in touch.