AI Engineering Roadmap

This is a self-contained roadmap page by an AI engineer (often referenced in ML communities) that lays out a structured learning path for becoming an "AI engineer" — someone who builds production systems around large language models (LLMs) and other foundation models, as opposed to a research scientist training new models from scratch. The field is evolving rapidly: traditional ML engineering focused on training custom models, but the 2022–2024 boom in APIs like OpenAI's GPT-4 and open-weight models (Llama, Mistral) shifted the job toward prompt engineering, retrieval-augmented generation (RAG), agentic workflows, and deploying pre-trained models. The page assumes the reader already knows basic Python and wants a curated, opinionated curriculum (it's not an official certification). Key prior context: "AI Engineering" as a distinct role solidified around 2023; prerequisites typically include Python, basic probability/statistics, and some familiarity with neural network concepts; the roadmap prioritizes practical deployment skills (APIs, vector databases, model serving) over deep learning theory. The author's perspective carries weight in self-taught and bootcamp-adjacent circles.