YYuhao Su·May 22The Live Index: Why Vector Search Should Be a Streaming ProblemVector databases assume the corpus is static. Real workloads aren't. Here's why vector search belongs in a streaming database, and how RisingWave's native vector type and streaming HNSW indexes close the gap between data that exists and data you can search.10 min read·risingwave.com
YYuhao Su·May 22GDPR-Compliant Streaming: Data Deletion in Real TimeArchitectural patterns for honoring the right to erasure inside a streaming database. How RisingWave propagates DELETE through materialized views, how upsert Kafka tombstones work, and how to design PII boundaries.9 min read·risingwave.com
RRisingWave Labs·May 22Stream Processing Without a Message Queue: Do You Really Need Kafka?Kafka is a reflex, not a requirement. Learn when a streaming database can ingest directly from databases, files, and HTTP, and when a message queue earns its keep.8 min read·risingwave.com
VVishal Uttam Mane·Apr 17Real-Time Data Analytics with Streaming Platforms This article explains how real-time data analytics leverages streaming platforms to process and analyze data instantly using distributed,4 min read·visha-uttam-mane-real-time-data.hashnode.dev
FFahad Shah·Apr 8AI Agent Data Infrastructure in 2026: What Your Agents Actually NeedRunning an LLM is now a commodity. The competitive moat is data: how fresh it is, how fast agents can access it, and how well it reflects current reality. Here is the infrastructure stack for 2026.18 min read·risingwave.com
FFahad Shah·Apr 8Choosing a Streaming Database in 2026: A When-to-Choose GuideRisingWave, Materialize, Apache Flink, ksqlDB and Spark Structured Streaming compared on what each does well, where each falls short, and when to pick it. No rankings.17 min read·risingwave.com
FFahad Shah·Apr 7Why AI Agents Need Real-Time Data (And How to Deliver It)AI agents make poor decisions on stale data. Learn how a streaming database delivers sub-second fresh context to your agents using plain SQL.11 min read·risingwave.com183A
FFahad Shah·Apr 7You Don't Need a Separate Vector DatabaseMost teams add Pinecone or Milvus on top of their existing stack, creating a second system to sync, pay for, and debug. If your data is already streaming, a streaming database with native vector support removes the extra layer.12 min read·risingwave.com
RRisingWave Labs·Apr 7Feature Backfilling: How to Generate Historical Features for Model RetrainingReconstructing historical feature values from raw events using the same SQL logic as your live pipeline — eliminating training-serving skew.11 min read·risingwave.com
RRisingWave Labs·Apr 7Replacing Your Feature Store: A Migration Guide for ML TeamsA phased approach to migrating from a dedicated feature platform to a streaming database — with an honest assessment of when it's worth it.12 min read·risingwave.com