Study library
Learn by building.
You leave able to implement ML from first principles, turn a paper into running code, and operate a long-context agent stack that is fast, cheap, and safe across tenants.
Today
This path is finished. Pick another, or wait for rust.
Going rusty
Your progress
Stored only in this browser.
0%
Problems
200 hands-on ML implementations
200 pagesResearch Papers
Implement landmark papers from scratch
172 pagesML Math
67 interactive math chapters
67 pagesLLM Internals
How modern LLMs actually work
7 pagesCS/AI 684
Efficient Long-Context Agent Systems (14 weeks)
20 pagesTensor→Tenant
108-week apprenticeship · 548 Darbar slots · 36-week on-ramp
716 pagesGuides
Writing & method
1 page