DataTalks.Club (Alexey Grigorev and Bhavani Ravi) · on GitHub
A free community course on building and shipping software with AI coding agents: spec-driven workflows with AGENTS.md, an AI-assisted full-stack app, testing and containerized deployment, observability, and agent building blocks like SKILL.md skills, subagents and parallel Git worktrees. It suits working developers who want an engineering-grade, project-based alternative to vibe-coding tutorials, with an optional cohort for accountability.
Aspiring builders, career switchers and working developers. This is an intermediate course in Coding with AI, running about 5 modules; 2026 cohort started 2026-08-31, self-paced anytime.
Free. Free; a certificate is available only to live-cohort learners who finish the final project and peer reviews, and self-paced learners are not eligible. Price seen 2026-11-01.
JetBrains' Paul Everitt shows how to evaluate an AI coding setup (the agent, the model, where it runs and how work is split) and rebuilds the same Python app in cloud, hybrid and fully local workflows. For developers weighing cost, privacy and capability trade-offs between hosted and local coding agents.
A feature-by-feature tour of Claude Code covering the agent loop, sessions, permission modes, sandboxing, hooks, headless runs, IDE and GitHub Actions integrations, CLAUDE.md memory, agent skills, MCP and multi-agent sub-agents. It suits developers who already use Claude Code casually and want one complete, structured reference instead of scattered short clips.
Taught by Anthropic's Elie Schoppik, it covers agentic coding practice in Claude Code: giving clear context, extending it with Playwright and Figma MCP servers, and applying it to a RAG chatbot codebase, a notebook-to-dashboard refactor and a web app built from a Figma mockup. For developers with Python and Git basics who want a project-driven complement to Anthropic Academy's Claude Code courses.