Recruiting: March 2026 Squad

Build Your Agentic Development Mech

Weeks / Fridays

Academic Hours

Max Participants

Yrs Exp Req

// Tactical Mech Display

Not AI slop — production-grade code with human oversight. In gaming, a mech is a powerful robotic vehicle piloted by a human. Our training is a virtual hangar where you construct your own — a collection of LLM loops, prompt libraries, and workflows that develop software semi-automatically. You are the engineer and pilot.

Join the Squad

Questions? Book a free strategy session

Registration closes February 21, 2026

Mech Assembly

6-Session Curriculum

Fridays (+ 1 Thursday) · March 6 – April 10, 2026 · 9:00–15:30 · via Google Meet

New to agentic coding? Learn what makes an AI agent different.

Fri, Mar 6

Core OS

Install your mech's operating system

Claude Code setup and IDE integration. Token economics and when to use extended thinking. Commands like /review, /test, /debug. Sub-agents for task delegation. Creating CLAUDE.md files that encode your team's conventions.

Fri, Mar 13

Context Engine

Teach your mech to understand intent

What makes agents understand your intent: code structure, documentation, and examples. Crafting Software Requirements Documents (SRDs) that capture scope, constraints, and acceptance criteria the agent can verify.

Fri, Mar 20

The Loop

Your mech's core operating cycle: think → act → observe

The Explore-Plan-Code-Commit pattern from Anthropic's engineering team. Breaking SRDs into agent-executable tasks. AI-enhanced TDD: write tests first, commit them, let Claude implement. Using separate Claude instances for independent code review.

Fri, Mar 27

Fleet Mode

Run multiple agents in parallel

Parallel development with git worktrees and multiple Claude instances. Agent debugging: logging, observability, and error feedback patterns. Feedback loops that let agents self-correct before commit. Generating C4 diagrams and Architecture Decision Records.

Thu, Apr 2 & Fri, Apr 10

First Deploy

Your mech's first real mission

Apply everything you've learned to build a real feature for your product. Work on your company's actual codebase with instructor support. Present your project and graduate with a production-ready deliverable.

Your Mech's Toolkit

Built in Your Own Codebase

CLAUDE.md — your mech's configuration file with team conventions, commands, and patterns
Prompt templates — reusable command modules for reviews, tests, and refactoring
SRD template — mission briefing format for converting requirements to executable specs
Git worktree setup — fleet launch pad for parallel Claude instances
Agent feedback loops — self-correction systems that verify before commit
C4 diagrams + ADRs — navigation charts for your project's architecture
Production feature — first deployment proof using the full agentic workflow

Requirements

Who Should Join

2+ years of professional software development experience
Comfortable with at least one programming language
Access to a codebase you can work with during training
Commitment to attend all 6 full-day sessions (9:00–15:30)
Curiosity to explore AI-assisted development workflows
Team player ready to share learnings with colleagues

Why This Works

Online tutorials don't stick. Developers watch videos, forget most of it, and keep coding the old way. This training uses your real codebase, so skills transfer immediately to daily work.

Live instruction from practitioners who use Claude Code daily — not recorded content from people who read the docs once.

Investment