The applied AI capability graph
Engineer the intelligence era.
Where AI engineers meet ambitious teams. Prompt design, retrieval, agent orchestration, evaluation, guardrails, and production hardening are different jobs. This directory makes the difference searchable — and asks every profile for proof.
What the index actually counts
Architectural integrity
We list capabilities and proof, not inflated headcount
Precision discovery
Filter talent through a capability taxonomy — prompting, retrieval, agents, evaluation — instead of the words “AI Engineer.”
Evidence on every profile
Published rows attach repos, write-ups, talks, or portfolios. Demo seed is flagged. Nothing is ranked or sold.
A shortlist that persists
Star profiles on Discover and compare them side-by-side. The list is stored on Convex for this browser session, with a local backup.
Latest capabilities
Technical verticals
Specialized work across the applied stack
Read the glossary →How it works
For employers
Define the stack
Name the capabilities you actually need — not a job title.
Review evidence
Open profiles and follow the proof links before you write.
Shortlist or write in
Star candidates onto a Convex shortlist, or submit a concierge brief.
For engineers
Name the capabilities
Pick the taxonomy slugs you practice. Leave the rest off.
Attach proof
A repo, a write-up, a talk. No score, no recruiter inbox.
Get found
Employers search the index by skill. Build a profile in four steps.
The index
- Andrej KarpathynanoGPT, llm.c, and Zero to Hero — language models built from first principles.2 evidence links
- Chip HuyenAuthor of Designing Machine Learning Systems and AI Engineering.2 evidence links
- Erik BernhardssonModal founder; Annoy and Luigi — vector search and pipeline plumbing.2 evidence links
Ready to look
Find the work, or add yours
Search the capability index, or paste your resume and review the profile it fills in.