The Complete 3-Day Curriculum
Eight hours of live instruction across three days. Every section, every topic, every outcome — right here, in full.
Three days, three transformations.
Day 1 builds the mental model and environment. Day 2 turns Claude Code into a real engineering tool. Day 3 takes you into agents, the API, and production-grade integration.
Day 01How AI Works + The Landscape + Setup
Foundations2 hours
How AI Works + The Landscape + Setup
Everyone starts Day 1 with a shared mental model of how LLMs actually think, a full map of the AI tooling landscape, a complete view of the Claude product suite, and a fully configured local environment.
AI Foundations
30 min- Brief history: ML → deep learning → transformers (2017) → GPT era → generative AI explosion
- How LLMs actually work: tokens, attention, context windows, temperature
- Why prompt engineering works — grounded in how the model thinks
- Key concepts: fine-tuning, RLHF, constitutional AI, system prompts
The AI Engineering Landscape
30 min- AI coding assistants: Cursor, Windsurf, GitHub Copilot, Codex, Replit Agent, Cline
- App builders / vibe coding: Lovable, v0, Bolt, Replit
- Where Claude fits versus all of the above
- AI workflow & automation: n8n, Zapier, Make
- Knowledge & productivity: Notion, Linear, and how MCP connects Claude to them
- Adjacent dev tools: Vercel AI SDK, LangChain, LlamaIndex
- Awareness plugs: ElevenLabs, Runway, HeyGen, Midjourney
- Models landscape: Claude, GPT, Gemini, Llama, Mistral
Claude: The Product Suite
30 min- Anthropic as a company
- Claude model family: Opus 4.6, Sonnet 4.6, Haiku 4.5
- The product map for engineers: Chat → Code → Agent SDK → Managed Agents → API
- Pricing: Pro, Max, API pay-as-you-go
Environment & IDE Setup
30 min- Claude Pro/Max account, Claude Code CLI, Desktop app, API key, Anthropic SDK
- IDE integration: VS Code + extension, Cursor, or terminal-only
- Clone the exercise repo, verify everything works
- Troubleshoot as a cohort — nobody leaves Day 1 broken
A shared language for how Claude works, the full ecosystem map, and a working local environment on every engineer's machine.
Day 02Claude for Engineering
Engineering3 hours
Claude for Engineering
We turn Claude Code into a real engineering environment — tuned to your conventions, connected to your tools via MCP, and wired up with custom commands. You leave Day 2 with a Claude-first repo on your own codebase.
Mastering Claude as an Engineer
45 min- The prompt engineering framework that actually moves the needle
- Model selection as a real engineering decision
- Extended thinking, Research mode, Projects, Artifacts, code execution
- Live patterns: debugging, refactoring, code review, ADRs
Claude Code: Your AI Engineering Environment
45 min- Core workflow and the permission model
- CLAUDE.md deep dive and AGENTS.md
- Memory and session management
- MCP servers: connect Claude to databases, APIs, GitHub, Notion, internal tools
- RAG concept: why context is everything
Building Your Personal Engineering System
45 min- Custom slash commands tuned to your workflow
- Skills system and event-driven hooks
- Git integration: branches, commits, PRs, reviews
- Turning a messy codebase into a Claude-first repo
Build Challenge
30 min- Hands-on: CLAUDE.md + MCP server + custom command
- Ship a real multi-step task end-to-end
- Live troubleshooting and review
- Leave with a working system on your own codebase
A Claude-first repo, MCP servers wired up, your own custom commands, and a reusable engineering workflow you can take to work tomorrow.
Day 03Advanced & Agentic Engineering
Agentic3 hours
Advanced & Agentic Engineering
The advanced day. Multi-agent orchestration, the full developer platform, the Agent SDK, Managed Agents, custom MCP servers — and a capstone project where you ship a real integration in your own app.
Multi-Agent Orchestration
45 min- Sub-agents and agent teams that actually hold together
- Git worktrees for parallel execution
- Headless mode, GitHub Actions, scheduled tasks
- Patterns for reviewers, implementers, and coordinators
The Claude Developer Platform & API
45 min- Messages API and client SDKs (TypeScript, Python, Go, Java, Ruby, PHP, C#)
- Tool use, function calling, structured outputs
- Prompt caching and the Batch API for cost control
- The Compaction API
- Production patterns: retries, streaming, observability
Agents at Scale: Agent SDK & Managed Agents
45 min- Agent SDK — Claude Code's engine, programmable in Python and TypeScript
- Managed Agents API: agents, environments, and sessions on Anthropic infrastructure
- The Advisor Tool pattern (executor + advisor model pairing)
- Building custom MCP servers with FastMCP
Capstone & What's Next
30 min- Full ecosystem map you now own
- Certificate of completion
- Alumni Discord + living curriculum
- 90-day plan: how to go from cohort to shipping
Multi-agent workflows running, a shipped API integration, and a capstone project you can show off on your first day back at work.
What you'll need on Day 1.
We want Day 1 to be real instruction, not setup. A few things will save us all time up front.
A working macOS laptop with admin access (to install tools)
Node.js 20+ installed (we'll verify in Day 1 setup)
A Claude Pro or Max subscription — or an Anthropic API key with a small budget for exercises
A code editor: VS Code, Cursor, or your preferred alternative
A repo you care about — ideally something you're actively working on. Exercises run against real code.
Everything in the cohort.
Secure your seat in the next cohort.
Waitlist members get early bird pricing ($399 vs $499). Cohorts are capped at 20 engineers and filling up.