ai-engineering
Read this series in order.
Each part builds on the one before it. Reading time, level, and the concepts covered are shown for every part below.
Intermediate
8 articles
108 minutes
8 concepts
Parts
Articles in this series
- Text to Tokens - The Foundation Text stops being text before the model ever sees it.
- Tokens to Embeddings - Vectors That Capture Meaning Meaning becomes geometry, then retrieval becomes possible.
- Embeddings to Attention - Relating Tokens to Each Other Every token asks which other tokens matter.
- Attention to Generation - Producing Text Token by Token Generation is a loop of probabilistic commitments.
- Generation to Retrieval - Grounding LLMs in Facts Hallucination starts where retrieval is absent.
- Retrieval to RAG - The Complete Pipeline RAG fails in the handoffs, not just the vector store.
- RAG to Agents - From Retrieval to Action Retrieval stops being enough when the model starts acting.
- Agents to Evaluation - Measuring What Matters Agents need evals before confidence means anything.