AI Without the Math PhD
AI & Machine Learning for Enterprise Architects
1. A working mental model: ML as systems engineering with a probabilistic component. 2. The seven core concepts — training, inference, embeddings, vector search, fine-tuning, RAG, MLOps — each translated into enterprise-architecture terms. 3. A reusable Rosetta Stone glossary mapping ML vocabulary to systems vocabulary. 4. The six EA lenses (requirements, interfaces, contracts, orchestration, pipelines, governance) applied to an ML system end to end. 5. A worked example — modernizing EDI with ML — so the bridge isn't theoretical. 6. A node-by-node learning path that uses an OSB Coach to translate any new concept into the systems language you already speak.
What This Book Will NOT Do
- It will not make you a data scientist or teach you to derive backpropagation.
- It will not bury you in math — the math exists, and you can learn it later if you ever need it. You mostly won't.
- It will not cover every algorithm. It covers the architecture that every algorithm plugs into.
- It will not pretend ML is "just software." The probabilistic part is real, and we'll name exactly where it bites.
Who This Is For
Enterprise architects, solution architects, integration architects, senior engineers, and technical leaders who already understand systems and now have to make decisions about AI. You can read a sequence diagram, you've governed a platform, and you're tired of AI explanations pitched at people who've never shipped anything.
Who This Is Not For (Yet)
If you've never designed or operated a real system, the analogies in this book are bridges to nowhere — you'd be translating into a language you don't speak yet. Start with a systems foundation first. And if your goal is to publish ML research, this is the wrong book; go get the math.
When This Isn't Enough