Case Study: Lens accelerates ML data processing with ClickHouse Cloud

How Lens made its database faster and more efficient

Lens, an open social protocol for blockchain-based social applications, needed a faster and more scalable data platform for its machine learning workflows. As the team’s data and feature needs grew, its existing Postgres and Rockset setup ran into ingestion delays, synchronization latency, and query concurrency limits, making real-time ranking and bot-detection tasks harder to support.

To solve this, Lens migrated to ClickHouse Cloud and used PeerDB to stream changes from Postgres into ClickHouse with a much simpler CDC pipeline. With ClickHouse, Lens cut table ingestion times from about 24 hours to just 1–2 hours, eliminated the latency caused by multiple data-transfer layers, and gained the ability to handle many concurrent queries without the bottlenecks they experienced before.

Lens