This post covers how teams building personalization, search or recommender systems can use Zipline to incorporate Semantic ID based generative recommendations into their stack.
A technical deep dive into how Zipline optimizes computation for cheap and fast batch jobs.
Streaming feature computation remains a challenge for many teams. This article goes over various architectures, their tradeoffs, and the approach that Zipline has found to work at scale.
AI platforms need to support a combination of read and write computed features/embeddings to effectively enable modern workloads.
This post covers the challenges of iterating on existing AI systems in a safe and efficient manner, and the approach that Zipline’s platform takes to solve these challenges.
