Learning paths
Walk the concepts, in order.
Eight curated paths through the 83 most-important AI and LLM concepts. Each path explains why every step matters and links to the full term entry.
From zero to LLM
The seven concepts that turn "AI" into a real mental model.
Start →22 min · intermediate · 7 termsThe attention family
Self-attention to FlashAttention — how the transformer's core operation evolved.
Start →24 min · intermediate · 9 termsThe post-training pipeline
Pretraining gives you a base model. This is how it becomes ChatGPT.
Start →30 min · intermediate · 13 termsThe RAG stack
From "let the model answer from documents" to a production retrieval system.
Start →28 min · intermediate · 13 termsHow modern teams actually evaluate LLMs
Past benchmarks: pairwise comparison, win rates, LLM-as-judge, online and offline.
Start →26 min · intermediate · 12 termsThe agent loop
Tool use, ReAct, plan-and-execute, MCP — how LLMs do work instead of just talking.
Start →26 min · advanced · 12 termsInference engineering
Make models fast and cheap to serve. The metrics, the techniques, the trade-offs.
Start →22 min · intermediate · 10 termsLLMOps & observability
Tracing every call, watching drift, feeding back into evals. Production AI discipline.
Start →How learning paths work
Each path is a hand-ordered sequence of glossary terms picked to build the right mental model. Read each entry top-to-bottom; click through to the full term page for the deep dive; come back to the path to continue. No accounts, no progress tracking — just a curated reading list.
Prefer free-range learning? The A–Z index lists every term alphabetically. Want to test yourself? Try the quiz.