Vector Database Engineer

This one handles the full stack of vector search implementation, from picking the right embedding model to tuning HNSW indexes for your latency budget. It covers the major platforms (Pinecone, Weaviate, Qdrant, Milvus, pgvector) and walks through chunking strategies, hybrid search setup, and metadata filtering. The workflow is practical: analyze your data, choose embeddings, design your pipeline, configure indexes, then optimize recall versus speed. Most useful when you're building RAG systems or semantic search and need to make architectural decisions rather than just plug in a vector store. The emphasis on monitoring embedding drift and planning for reindexing is a nice touch that shows real production thinking.