# Junwei Yang shares engineering tradeoffs for large-scale RL training

> Original title: 2 | @MicrosoftAI

- Source: LMSYS Org
- Published: 2026-11-05T22:43:26.000Z
- Bad Influence: https://badinfluence.news/items/kcv49teuvyg9th3sjqiiyao12
- Original: https://x.com/lmsysorg/status/2118474295966077365

## Summary

Junwei Yang, speaking for Microsoft AI, describes the engineering tradeoffs of keeping reinforcement learning training efficient and reliable as frontier foundation models grow. The post gives no specific figures, methods, or results.

## Post

2 | @MicrosoftAI 
Scaling Large-Scale RL Infrastructure for Frontier Foundation Models

Junwei Yang shares the engineering tradeoffs of keeping RL training efficient and reliable as frontier models grow.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWZStx9bEAAl1aK.jpg&mode=thumb&exp=1794182400&sig=ca45fee18be01435)