Build AI agents, architect the platform, make it auditable — a free, bilingual course

AI Architect Trainer is an independent, free training platform for the people who put AI systems into production: engineers who build agents on large language models, architects who design the data platform those agents run inside, and the governance and audit specialists who have to prove the result is under control. Every lesson, question and flashcard exists in both Ukrainian and English. There is no paid tier.

Why this site exists

Most AI training teaches one of these in isolation: a prompting course, a data-platform course, a compliance course, a certification cram. The work does not arrive in pieces. An agent that reaches production needs a platform beneath it, access rules around it, lineage behind every answer it gives, and an AI management system an ISO/IEC 42001 auditor can test. This site teaches those disciplines as one sequence — build, architect, govern, certify — so each one is learned next to the others it depends on.

Five disciplines, one arc

Build: Claude and AI agent engineering (L1–L14)

The working surface of the Claude ecosystem: the Messages API, tool use and tool design, Claude Code, the Model Context Protocol, the Agent SDK, multi-agent patterns, extended thinking, computer use, deployment on Bedrock and Vertex, prompt caching, the batch API and evaluation harnesses. Every module ships lessons, cheat sheets, spaced-repetition flashcards, interactive simulators and hands-on labs.

Architect: enterprise AI architecture (E1–E8)

The platform a model runs inside, across four layers: the data fabric and distributed compute beneath it, lakehouse storage and incremental views, the semantic ontology and domain model that give data meaning, entity resolution and knowledge graphs, attribute-based access control and policy enforcement, provenance and lineage for compliance, decision intelligence, and consensus between reasoning agents. Each lesson names the peer-reviewed papers it derives from.

Govern: AI audit and governance with ISO/IEC 42001, the EU AI Act and IAPP AIGP (I1–I8, G1–G13)

How an AI management system is audited: its foundations, the normative requirements of Clauses 4–10, Annex A and the Statement of Applicability, audit method under ISO 19011 and ISO/IEC 17021-1, gathering objective evidence on site, writing and closing nonconformities, and mapping the management system onto the obligations of the EU AI Act. Worked audit cases, auditor working templates, and a full timed simulation of the lead auditor exam in the PECB format. Independent material, not affiliated with ISO, IEC or PECB.

IAPP AIGP (G1–G13): preparation for the IAPP Artificial Intelligence Governance Professional exam on Body of Knowledge v2.1 — the foundations of AI governance; laws, standards and frameworks, including the EU AI Act, the GDPR, the NIST AI RMF, ISO/IEC 42001 and US law; and governing AI development, deployment and use. A question bank with case studies and multi-select items, and an exam simulation built on the BoK blueprint. Independent material, not affiliated with or endorsed by the IAPP.

Certify: CCA-F exam preparation

Independent study material for Anthropic's Claude Certified Architect — Foundations exam, mapped to the five published domains: agentic architecture and orchestration, tool design and MCP integration, Claude Code configuration and workflows, prompt engineering and structured output, and context management and reliability. Scenario questions, a timed mock exam and a weak-spot review loop.

Start here: AI for everyday work

For people arriving without a background in this: everyday workflows, prompt patterns that generalise, artifacts, and an honest look at what tokens cost.

Guides

AI Architect Trainer is an independent, community-made study aid. It is not affiliated with, authorized, or endorsed by Anthropic. “Claude” and “Claude Certified Architect” are trademarks of Anthropic. Exam details are based on Anthropic's published certification blueprint and may change — always confirm current details on Anthropic's official certification pages.