A personal quest to master SysML, now with an LLM librarian

What it does

This is a curated learning repository for SysML — the systems side of machine learning (think distributed training, serving, and the infrastructure underneath ML). It bundles a road map, papers, courses, videos, open-source frameworks, and a glossary of basic ML concepts. The author publicly committed to becoming a SysML expert in 3–5 years and records every step along the way.

The interesting bit

The repo recently adopted a three-layer “LLM-maintained wiki” workflow inspired by Karpathy: raw/ holds immutable source material, wiki/ holds agent-maintained cross-linked synthesis, and AGENTS.md defines the ingest, query, and lint workflows. So it’s part study plan, part experiment in letting agents do the note-taking.

Key highlights

• A road map (map.md) organizing papers, courses, videos, and frameworks into one path
• concepts.md collects basic ML definitions in one place
• Agent-maintained wiki layer with a compiled index at wiki/index.md
• Existing topic directories stay intact and migrate incrementally
• The author invites you to watch the repo for updates as it grows

Caveats

• It’s a personal learning notebook, not a library or tool — depth and coverage depend on wherever the author happens to be on the journey
• The wiki layer is a work in progress; how well the agent synthesis holds up is unclear from the README

Verdict

Worth a look if you want a structured, opinionated starting point for learning ML systems — or a real-world example of the LLM-wiki pattern. Skip it if you expected a framework or runnable code.