From the source
It was energising to see the community come together at Epoch with such a strong sense of possibility.
The Builder Edition brought together developers and researchers for technical sessions and live demos.
The Enterprise Edition brought together leaders from banking, government, and industry to explore what it takes to deploy AI across large organisations.
Across the two days, we shipped across the entire stack: model updates, an inference layer hosted in India, a self-serve agent platform, and partnerships designed to carry these capabilities into institutions across the country.
This post covers everything we announced.
Models and products Sarvam 105B We released the first Sarvam 105B checkpoint in February.
At Epoch, we released updated variants for conversational agents and work agents, the two use cases where we are seeing the most traction in India.
For conversational agents, standard public benchmarks are incomplete.
They do not capture ambiguous messages, ASR noise, code-mixed speech, long policy documents, or instructions that must be followed many turns later.
…
