Course Wiki

AI Dev Tools Zoomcamp

DataTalks.Club's free course on AI-native software engineering: coding agents, context engineering, full-stack builds, CI/CD, observability, MCP, and reusable agent skills.

AI Dev Tools Zoomcamp is DataTalks.Club’s free, hands-on course on disciplined AI-assisted software development. It is built around one workflow: give AI tools the right context, use them for the right job, review what they produce, test the result, and ship software with guardrails. The course compares modern AI developer tools, builds and deploys a full-stack application, operates it with observability and agent-assisted incident response, and extends coding agents with MCP and reusable capabilities.

This is deliberately not a prompt-engineering course and not a model-training or RAG course — the focus is day-to-day software development with AI coding assistants, agents, tests, CI/CD, deployment, documentation, and review. All materials are open source in the course repository, with videos on YouTube and a course FAQ. The 2026 cohort starts August 31; materials are also usable self-paced.

Curriculum

All modules, each with its lesson-level notes:

• Module 1: AI-Native Developer Workflow
• Module 2: Build and Ship an AI-Assisted Full-Stack App
• Module 3: Test, Containerize, and Deploy an AI-Assisted App
• Module 4: DevOps and Observability for AI-Built Apps

• Module 5: Coding Agent Capabilities

• Module 1: AI-Native Developer Workflow — turn a vague idea into a spec and a task backlog (spec-driven development), give coding agents durable context through AGENTS.md (context engineering), and run PM/engineer/QA roles with loop and graph engineering (coding agents).
• Module 2: Build and Ship an AI-Assisted Full-Stack App — frontend prototype, an OpenAPI contract, a FastAPI backend with auth and real-time collaboration, SQLite persistence, and tests.
• Module 3: Test, Containerize, and Deploy — Docker Compose with Postgres, integration and end-to-end tests including Playwright, AWS deployment, and a CI/CD pipeline that deploys only after tests pass.
• Module 4: DevOps and Observability for AI-Built Apps — separate dev and prod environments with a promoted release pipeline, OpenTelemetry into Prometheus, Loki, Tempo, and Grafana, actionable alerts, and a bounded, read-only coding-agent responder that investigates incidents.
• Module 5: Coding Agent Capabilities — MCP servers, discoverable agent skills with SKILL.md, focused subagents that separate implementation from independent review, and parallel task orchestration with isolated Git worktrees.

The final project applies the workflow end to end: an app of the learner’s own with frontend, backend, an API contract, persistence, tests, containerization, a public deployment, and documentation of exactly how AI tools, prompts, and agent instructions were used.

Looking for a specific concept? The Concept Glossary links every concept this course teaches to the exact modules and lessons that cover it.

Who it is for

The course targets people who already write basic code — software engineers, ML and MLOps engineers, data scientists, data engineers, and analysts who code — and want to use AI coding tools professionally with a repeatable workflow for planning, prompting, reviewing, testing, and shipping. No prior AI tool experience, powerful machine, or GPU is required. Certificates require the final project plus the required peer reviews during a live cohort.