Embeddings & Vector Search

Understand how text becomes vectors and how to search them: similarity measures, learned embeddings, exact and approximate nearest neighbours, quantisation, hybrid search, evaluation and production, with every experiment run for real.

Start course →

Syllabus

What Embeddings Are

  1. From Words to Coordinates
  2. Where Embeddings Come From
  3. Similarity Measures: Dot, Cosine and Euclidean
  4. Dimensions and the Curse of Dimensionality

Creating and Choosing Embeddings

  1. A Learned Embedding You Can Inspect
  2. Keyword Search vs Semantic Search
  3. Choosing an Embedding Model
  4. Preparing Text: Chunking, Truncation, Query vs Document

Searching Vectors: Exact and Approximate

  1. Exact k-Nearest Neighbours
  2. Approximate Search: IVF and HNSW
  3. Compression: Scalar and Product Quantisation
  4. Choosing an Index

Beyond Plain Nearest Neighbours

  1. Metadata Filtering
  2. Hybrid Search and Reranking
  3. Clustering, Topics and De-duplication
  4. Multi-Vector, Multilingual and Multimodal Embeddings

Evaluating and Improving Retrieval

  1. Recall@k, MRR and nDCG on a Golden Set
  2. Error Analysis and Improving Quality
  3. Model Changes, Re-Indexing and Drift

Running Vector Search in Production

  1. Where Vectors Live
  2. Updates, Deletes and Freshness
  3. Scaling, Cost and Latency
  4. Security and Privacy of Embeddings

Applications of Embeddings

  1. Semantic Search and Recommendations
  2. Classification, Routing and Few-Shot Labelling
  3. Embeddings for RAG and Agent Memory

Putting It Together

  1. Case Study: Semantic Search for a Help Centre
  2. Revision: Cheat Sheet and Self-Check