Phase 3 · Embeddings & RAG

Ground answers in your own data — embeddings and retrieval with a vector database and a metadata store

🔍 Phase 3 · Embeddings & RAG9 lessons~217 min total reading

Lessons in this phase

Intermediate~16 min

#21Embedding Models: Representations, Similarity & Fine-tuning

Intermediate~26 min

#22Vector Databases: Indexing, ANN Search & Production Patterns

Intermediate~30 min

#23LanceDB: Embedded Multimodal Vector Lakehouse

Intermediate~15 min

#24Chunking Strategies: Splitting, Overlap & Semantic Boundaries

Intermediate~30 min

#25Retrieval Strategies: Hybrid Search, Reranking & HyDE

Intermediate~31 min

#26RAG: Retrieval-Augmented Generation as an Agentic Pattern

Intermediate~5 min

#27Advanced RAG: Agentic, Graph-Based & Multi-Hop Retrieval

Intermediate~31 min

#28RAG Evaluation: Faithfulness, Relevance & Failure Modes

Intermediate~33 min

References & further reading

What are embeddings?Cloudflare
Plain-English primer: how embeddings turn text into vectors that capture semantic meaning.
EmbeddingsAI SDK (Vercel)
Generate embeddings with embed / embedMany and compare them via cosine similarity in TypeScript.
Vector embeddingsOpenAI
Use embeddings for semantic search, clustering, recommendations, and classification.
EmbeddingsGoogle ML Crash Course
Why lower-dimensional learned representations beat sparse one-hot encodings.
What is embedding?IBM
Conceptual overview of embeddings and their role across modern ML workloads.
EmbeddingsGoogle Gemini API
Generate text and multimodal embeddings through the Gemini API.