Prerequisites: Solid understanding of LLMs, embedding models, foundational vector search concepts, and experience with Python programming and machine learning libraries.

Level: Advanced

What You'll Learn

• Advanced Indexing Algorithms

  Implement and analyze complex Approximate Nearest Neighbor (ANN) index structures like HNSW, IVF variations, and graph-based methods.

• Vector Search Optimization

  Apply techniques such as quantization (scalar, product), advanced filtering, and hardware considerations to optimize search latency and throughput.

• Hybrid Search Implementation

  Combine vector search with traditional methods like keyword search (BM25) and graph techniques for improved relevance.

• Metadata Filtering Strategies

  Design and implement efficient pre-filtering and post-filtering mechanisms using associated metadata.

• Scalability and Deployment

  Architect and manage distributed vector search systems for large-scale, production LLM applications.

• Performance Tuning

  Tune index parameters and query execution plans for optimal recall, precision, and speed trade-offs.