parsec

save 50% of your coding agent limit.

[surya @subygan]

run gemma-4B in a Jev compatible API with parallel decoding. get responses in 100ms

with a finetunable head this should get much better results on specific tasks

https://t.co/YfEa5zTBXy

[surya @subygan]

setting up the booth at the AI infra summit!