# Decagon uses SGLang in an AutoResearch loop for inference optimization

> Original title: Love seeing SGLang in @DecagonAI's inference stack! The team has built an AutoResearch loop where agents diagnose bottlenecks, test chang...

- Source: SGLang
- Published: 2026-10-29T00:29:44.000Z
- Bad Influence: https://badinfluence.news/items/c5dg7vrjcxok3f506810mf60s
- Original: https://x.com/sgl_project/status/2115601944412852607

## Summary

SGLang's account says Decagon runs its inference stack on SGLang and has built an AutoResearch loop. In the loop, agents diagnose bottlenecks and test changes against production traffic replays before building on verified gains.

## Post

Love seeing SGLang in @DecagonAI's inference stack! The team has built an AutoResearch loop where agents diagnose bottlenecks, test changes against production traffic replays, and build on verified gains. 

Check out how it's done 👇

## Quoted post @DecagonAI

> https://x.com/i/article/2115502716361560064

https://x.com/DecagonAI/status/2115511110120120437