---
title: "Krishna Prasad Srinivasan"
category: "people"
role: "Head of Vision Models"
company: "Sarvam"
linkedin: "https://www.linkedin.com/in/krishnapsrinivasan/"
twitter: "https://x.com/fewshotlearner"
sourceLabels: ["Official speaker roster", "Official conference schedule"]
---
# Krishna Prasad Srinivasan

## Profile
Head of Vision Models at [[sarvam|Sarvam]].

- [LinkedIn](https://www.linkedin.com/in/krishnapsrinivasan/)
- [X / Twitter](https://x.com/fewshotlearner)

## Biography
Krishna Prasad Srinivasan is a Head of Vision Models at [[sarvam|Sarvam]], where he led a lean team to train [[sarvam|Sarvam]] Vision, India's first sovereign VLM: a 3B state-space model that topped global OCR benchmarks at launch and led the Indic OCR Bench across 22 languages. He now leads the vision vertical's models, research, and product. Previously, he was Tech Lead for AI at Microsoft Research, where he built multilingual copilots for education and developed Indic translation models that outperformed commercial systems. Before that, Krishna was a researcher at Harvard, where he engineered a novel OCR architecture using contrastive learning that outperformed industry benchmarks on complex multilingual documents.

## Conference Sessions
- [[2026-06-29-krishna-prasad-srinivasan-from-scratch-to-sota-training-a-3b-state-space-vision-model-for-1-4-billion-people]] — From Scratch to SOTA: Training a 3B State-Space Vision Model for 1.4 Billion People (2026-06-29, 3:20pm-3:40pm)

## Evidence Graph
This evidence graph summarizes how this person appears across the conference source graph: scheduled sessions, linked videos, transcripts, and slide-derived evidence.

### Linked Sessions
- [[2026-06-29-krishna-prasad-srinivasan-from-scratch-to-sota-training-a-3b-state-space-vision-model-for-1-4-billion-people|From Scratch to SOTA: Training a 3B State-Space Vision Model for 1.4 Billion People]]

### Media Signals
No linked video, transcript, or slide source has been attached yet.