Thanks for the link! I am not familiar with the company but reminds me of the whole formal methods debate in distributed systems. Sure, writing TLA+ specs is the 'correct' deterministic way to build a Raft implementation, but in reality everyone just writes messy Go/Java and patches bugs as they pop up because its faster.

Was with you up to

> because its faster

Interesting, did you have any learnings that would apply to this problem now?

Check out our blog series here: ucbskyadrs.github.io!

As described in our VLDB paper (https://www.vldb.org/pvldb/vol17/p2694-cheng.pdf), we had to implement a number of optimizations and integrate scheduling with CC to get good performance.

What did you use to try to optimize scheduling?

Awesome results! Did you use the simulator from the NSDI paper directly?

Neat, thanks for the link! Seems like the agent feedback loop made a big difference for your use case. Would be interesting to see how much of a difference more advanced methods (like the genetic algs you mentioned) would make

The algorithm works for MoE load balancing in general

We've found that these frameworks do well for systems performance problems and expect that the range of problems for which they apply will increase as they models and frameworks improve. See our paper (https://arxiv.org/pdf/2510.06189) for more discussion about this

That's a good point! The load balancing of the original algorithm was already quite good so our goal was to try to get something that could achieve similar results but could run faster since runtime was also a concern.