Tier 1: Foundations — About This Tier
Tier 1: Foundations
Section titled “Tier 1: Foundations”Overview
Section titled “Overview”Tier 1 is designed for everyone: developers, DevOps engineers, product owners, and technical leads. This tier covers the theoretical foundations of AI-augmented development and gets the entire team hands-on productive with Claude Code in a single day.
This is not a “product owner lite” track—product owners derive genuine, immediate value from Tier 1 without needing Tier 2. You’ll understand how LLMs actually work, learn the prompt engineering techniques that separate productive developers from frustrated ones, and master the workflow inversion where clear specifications become the critical path. For technical staff, you’ll install and operate Claude Code, set up your first agent integrations, and learn the context engineering techniques that determine whether you complete a day’s work in an afternoon or spend 8 hours fighting the tool.
Duration: ~7 hours total (5 modules, 45–90 min each) Audience: Developers, DevOps engineers, product owners, technical leads Prerequisites: None. Basic familiarity with your development environment helpful but not required.
The 5 Modules
Section titled “The 5 Modules”M01: How LLMs Actually Work (and Why It Matters for Your Prompts)
Section titled “M01: How LLMs Actually Work (and Why It Matters for Your Prompts)”Understand transformer architecture, attention, token-by-token generation, and why hallucinations happen. Install Claude Code and run your first real interaction with a live codebase.
M02: Prompt Engineering — From Vague Requests to Precise Instructions
Section titled “M02: Prompt Engineering — From Vague Requests to Precise Instructions”Learn zero-shot, few-shot, chain-of-thought, and the four proven principles that separate 80% time savings from frustration. Pick the right model for the job.
M03: Specifications Are the New Source Code
Section titled “M03: Specifications Are the New Source Code”Master Plan Mode. Shift the workflow from chaotic “try it and see” to idea → prototype → crystal-clear spec → execution. This module is especially critical for product owners and project leads.
M04: Context Engineering — The Only Lever That Matters
Section titled “M04: Context Engineering — The Only Lever That Matters”Prevent the four failure modes that kill productivity in large codebases: poisoning, distraction, confusion, and clash. Build and maintain your CLAUDE.md, understand token accounting, and learn when to segment work across multiple sessions.
M05: Agents and MCP — What They Are, How They Reason, Why They Fail
Section titled “M05: Agents and MCP — What They Are, How They Reason, Why They Fail”Connect your first MCP servers (GitHub, databases, team tools). Understand why agents sometimes make bad tool choices and how to curate your integrations for reliability.
What You’ll Have at the End
Section titled “What You’ll Have at the End”- Claude Code installed and configured on a real project
- A personal prompting cheat sheet and habit loop
- A team specification template and Plan Mode workflow
- A committed CLAUDE.md with project conventions
- Working agent integrations (MCP servers) relevant to your team
- A mental model of LLM strengths and weaknesses that informs every future interaction
Estimated time investment: 7 hours over 1–2 days Expected outcome: Immediate productivity gains (50–80% time savings on routine tasks), confidence to expand into Tier 2