@lindavivah: Everyone's building agents. So why did Redis release a feature store? I went on a tech walk with Simba Khadder, founder of Featureform (acquired by Redis) & now Head of Context Engine at Redis, to ask him exactly that and he dropped some fantastic insights! "It's almost like everyone forgot - some of the most important models we have are still classical ML. And those models need features."
 He also shares the feature behind one of his most meaningful wins predicting subscribers at 100M+ users, in the work that led him to build Featureform. The bigger picture: models need the right data at the right moment. So do agents. Models call it features, agents call it context, and Redis is building a real-time data layer under both.

Feature Form's core idea: Define your features as code. They become versioned, reusable pipelines you write once and use for both training and inference, and it works with the stack you already have (Spark, Snowflake, Kafka, Redis as the online store) instead of replacing it. Shout out to Simba & Redis for taking the time to tech walk & talk in SF! #techtok #machinelearning #EduTok

lindavivah

Open In TikTok:

Region: US

Wednesday 29 July 2026 18:59:05 GMT