@ShubhamMal72313 Ha, Iโd take the GPUs too ๐ Read TrnRT post, NKI megakernels are solid work, and I respect that you published the caveats. Around KV cache the hard part is keeping the hit rate up when many tenants share the tier and a fetch has to beat recompute. Happy to connect.
Gpu memory forgets an agent's prompt in about 4 seconds
agents return from tool calls much later, so nearly every turn re-read the whole prompt
a shared cache tier that holds prompts for hours halved the recompute, across 450B+ tokens a day
Sarvam token factory is GA ๐
Today, we are launching Sarvam Token Factory for everyone.
We host our own models, and the full range of best-in-class open-weight models, in India. We match leading inference providers on performance and reliability.
For enterprises, this means more models to choose from, and an AI stack they can own without lock-ins.