A series is a path with prerequisites and an outcome. Categories remain a separate way to browse topics without implying reading order.

An agent engineering learning path In progress

Learn Agents from 0 to 1

Build the right mental model through short lessons, then use Pi to ship an evidence-first technical research agent.

2

published

28

lessons

7

stages

1. 01See the map first
2. 02Understand the LLM runtime
3. 03Give the agent hands
4. 04Manage context and memory
5. 05Learn agent patterns
6. 06Move from demo to reliable system
7. 07Ship the complete agent project

Open learning path →

A Pi Agent implementation path Planned

Build a Pi Agent from 0 to 1

Start with the smallest runtime loop, then build a runnable, testable, and extensible Pi Agent step by step.

Roadmap follows after the current series is finalized

PromptWeave architecture series In progress

Designing Prompt Systems: From Text to Rollbackable Runtime Versions

Start with design scenarios where prompts, models, and experiments become coupled, then derive versioning, routing, fallback, and release snapshots.

0

published

4

lessons

3

stages

1. 01Establish the problem boundary
2. 02Protect one execution
3. 03Ship the verified version

Open learning path →