Hybrid Search and Reranking for Production RAG: An ADR
Hybrid BM25 plus dense retrieval with reranking: architecture, fusion methods, latency and cost, and when to skip reranking. A 2026 decision record.
Flink 2.2 ML_PREDICT and VECTOR_SEARCH: Running Model Inference Inside a Stream
Flink 2.2 adds VECTOR_SEARCH and Table API ML_PREDICT for in-stream inference. How they work, where they break, and when to keep inference outside.
pgvector vs Qdrant vs LanceDB: On-Prem RAG Vector Search (2026)
pgvector, Qdrant, and LanceDB compared for on-prem RAG in 2026: indexing, filtering, latency at scale, and operational overhead.
ColPali and Visual Document Retrieval: Late-Interaction RAG (2026)
ColPali explained: how late-interaction vision-language retrieval indexes document page images directly - MaxSim scoring, storage trade-offs, and a production multimodal RAG architecture.
Matryoshka Embeddings: Adaptive-Dimension Retrieval Architecture (2026)
Matryoshka embeddings explained: how Matryoshka Representation Learning nests multiple dimensions in one vector for adaptive retrieval - coarse-to-fine search, storage cuts, MRL training and failure modes in 2026.
Hybrid Search Architecture: Dense + Sparse Fusion with RRF (2026)
Hybrid search architecture explained: fusing BM25 sparse retrieval with dense vector search using Reciprocal Rank Fusion - indexing, scoring, rerankers, latency and failure modes for production RAG in 2026.
LLM Semantic Caching Architecture: Cut Inference Cost and Latency (2026)
A semantic caching architecture for LLM apps: exact vs embedding-similarity cache tiers, thresholds, invalidation, eviction, and the cost/latency math behind GPTCache-class systems.
Embedding Models Benchmark: OpenAI, Cohere, Voyage, BGE
A 2026 embedding models benchmark: OpenAI, Cohere, Voyage, and BGE on retrieval quality, dimensions, cost, and MTEB - with what changed for 2026.
Vector Search in CouchDB: Options & 2026 Alternatives
Vector search in CouchDB in 2026: what is native vs not, integration patterns with dedicated vector databases, hybrid search, and when to migrate.