PM Mechanics

Public, read-only mirror of the pm-mechanics curriculum — browse the entire material without cloning the repo. This site rebuilds automatically from main on every push.

A mechanics trainer, not a desk-grade risk system: no live market data, no production curve or optimization engine. The unit of work is a formula made concrete, unit-aware, and tested.

For the interactive version — running the tests, /pmexpert, /tutor, /pm-query — clone the repo instead; see Getting started.

Where to start

• Repository overview — what's here, how deep it goes, and what's still conceptual-only
• Mastery guide — how the interactive /pmexpert trainer works (only usable once you've cloned the repo)
• Learning paths — choose a route by goal
• Reference index — every concept and instrument page, by topic
• Curriculum — the 14-day bootcamp, day by day
• Use cases — realistic portfolio workflows

There's also a lighter, single-page Artifact preview covering the reference pages, curriculum, and use cases — useful if this site isn't reachable. It's refreshed manually rather than on every push; this site is the always-current one.

Related repositories

This repo is the math layer: the deterministic analytics and the curriculum that teaches you to derive them. agentic-pm-lab is the agent layer built on the same domain — how to let an LLM call analytics like these without it inventing the numbers, using LangGraph/Deep Agents, Cedar authorization, evaluation, OpenTelemetry, MCP, and AWS Bedrock AgentCore.

Read them in that order. This repo answers "what is DV01, and how do I compute it correctly?"; agentic-pm-lab answers "how do I put a DV01 tool behind an agent and still trust the answer?" See ROADMAP's Phase 12 for where the two meet.