Level 2 · Intermediate — Primwright: Applied AI Systems
An Intermediate Guide to Building, Connecting, Evaluating, and Securing AI Applications

Book 2 · Applied AI Systems
Book 2 of 6 · Level 02 · Intermediate · 02 · Assembly
An Intermediate Guide to Building, Connecting, Evaluating, and Securing AI Applications
The second level of the Primwright progression — for readers who have completed Foundations and are ready to build with AI: connecting systems, evaluating behavior, and securing what they ship.
Coming soon
What the level covers.
- Translate a user need into requirements, boundaries, acceptance tests, and a fallback.
- Version prompts and context as application components.
- Validate syntax, schema, and business rules.
- Read and modify a small Python/API workflow.
- Build semantic search and a small RAG system.
- Define narrow tools, approvals, and safe stops.
- Implement or simulate one bounded agent loop.
- Create evaluation datasets, rubrics, and regression tests.
- Threat-model injection, exposure, and unsafe tool use.
- Instrument a prototype with useful logs and traces.
- Present evidence, limitations, residual risk, and next steps.
Pillar 01 · Building. Applying the judgment from Foundations to systems you construct — not just tools you use.
Pillar 02 · Connecting. Joining AI capabilities to real workflows, data, and software — the seams where most failures live.
Pillar 03 · Evaluating. Measuring what your systems actually do — with rubrics, tests, and documented limits.
Pillar 04 · Securing. Protecting what you ship — the practical security posture of AI applications.
The units, in order.
- Unit 1 · From AI User to AI Builder
- Unit 2 · Advanced Prompting and Context Engineering
- Unit 3 · Structured Outputs and Reliable AI
- Unit 4 · APIs and AI Application Fundamentals
- Unit 5 · Embeddings and Semantic Search
- Unit 6 · Retrieval-Augmented Generation
- Unit 7 · Tool Calling and AI Automation
- Unit 8 · Agents and Agentic Workflows
- Unit 9 · AI Evaluation and Quality Engineering
- Unit 10 · AI Security, Privacy and Responsible Use
- Unit 11 · Production-Style Prototype Discipline
Who it's for.
Readers who have completed Foundations — who can understand, use, and verify AI — and are ready for the next step: applying that judgment to real systems they build, connect, evaluate, and secure.
Prerequisites.
- Book 1 (Foundations) — each level assumes the last; none can be skipped.
- Not required: Python, command-line fluency, paid subscriptions, prior API use, advanced mathematics, or model-building experience.
What you get.
Every level pairs a book with a course. The book is the curriculum — the complete text and the standard of evidence. The course is the guided path through it — sessions and knowledge checks.
The book. Primwright: Applied AI Systems — intermediate guide to building, connecting, evaluating, and securing ai applications at Level 02 (Intermediate).
The course. Course 2 is the guided companion to Book 2 — sessions and knowledge checks. Format and access are announced with the release.
Availability, stated plainly.
Coming soon · Format, length, price, and release date will be announced — nothing is for sale yet, and no dates are promised.
After Level 02: AI Engineering & Production Systems.
Level 03 · Advanced — An Advanced Guide to Engineering, Securing, Deploying, Observing, and Optimizing AI Systems
Level 3 · Advanced — Primwright: AI Engineering & Production Systems