[ blog ]

Field notes from running models at scale

Engineering deep-dives, product updates, and notes from the team.

What Changes Between 100 Million Vectors and 12 Billion
Engineering

What Changes Between 100 Million Vectors and 12 Billion

Nothing about the query changes. Everything about memory, rebuilds and the shape of your on-call does.

Read more
Engineering

How We Cut p99 by 85% With Adaptive Batching

Read more
How We Cut p99 by 85% With Adaptive Batching
Engineering

A Practical Guide to Vector Search at Production Scale

Read more
A Practical Guide to Vector Search at Production Scale
Product

Designing a Recall Report Engineers Actually Read

Read more
Designing a Recall Report Engineers Actually Read
Engineering

Moving Off pgvector Without a Migration Weekend

Read more
Moving Off pgvector Without a Migration Weekend
Engineering

Inside Nodeform: Our Multi-Region Failover Architecture

Read more
Inside Nodeform: Our Multi-Region Failover Architecture
Open Source

Why We Open-Sourced Our Recall Harness

Read more
Why We Open-Sourced Our Recall Harness