🚨🚨🚨Excited to share our latest work: "Pralekha: An Indic Document Alignment Evaluation Benchmark", focusing on document-level alignment across 11 Indic languages. Paper: arxiv.org/abs/2411.19096 Github: github.com/AI4Bharat/Pral… Hugging Face 🤗: huggingface.co/datasets/ai4bh… 🧵1/N

Introducing Indic Parler-TTS: Open-Source Text-to-Speech for Over a Billion Indic Speakers! 🌏 In collaboration with Hugging Face, we are excited to release Indic Parler-TTS, a state-of-the-art open-source text-to-speech system designed to bring accessible and high-quality

#NLProc Just because GPT-4o is 17 times more expensive than GPT-4o-mini, does that mean it generates synthetic data 17 times better? Introducing the AgoraBench, a benchmark for evaluating data generation capabilities of LMs.

🚨🚨🚨 Tutorial accepted to EMNLP 2025! Anoop Kunchukuttan, Rudra, Mohammed Safi Ur Rahman Khan, Thanmay and I will be doing a tutorial on "Data and Model Centric Approaches for Expansion of Large Language Models to New Languages" at EMNLP. You want to expand LLMs? We will share our knowledge on how!

I'll be presenting our poster for IndicVoices-R at NeurIPS Conference (Friday, Dec 13, 11 AM, West Ballroom A-D #5110) I'd love to chat about the work we do at AI4Bharat, from speech to LLMs! 📚I’m also seeking summer research internships, would appreciate any guidance or connections

Glad to share we have a tutorial accepted at EMNLP 2025 on "Data and Model Centric Approaches for Expansion of Large Language Models to New Languages" - with @prajdabre1 Mohammed Safi Ur Rahman Khan Thanmay Rudra Murthy You can see an early version of this here: anoopkunchukuttan.gitlab.io/publications/p…

On the left we have a billion dollar company with restrictive license, On the right we have a open source non profit company with MIT license. They blatantly copied; they didn't even have the courtesy to clone the repo and make changes.

Sarvam is dedicated to building the bedrock of Sovereign AI for India. We see immense value in creating a Sovereign AI ecosystem in India, not to decouple from the world but to ensure strategic autonomy in building foundational components tailored to India's unique needs. We

Presenting the IndicRag Suite for up to 19 Indian languages! A LLM can't possibly have all the information in real time and it can hallucinate from time to time. Grounding it in context is the best possible way to get it to answer satisfactorily. LLMs for Indian languaged suck

Introducing Cadence, our new SOTA generalist punctuation restoration model supporting all 22 scheduled languages of India, along with English. Cadence is designed to "Mark My Words" accurately, handling both clean written text AND messy, spontaneous spoken transcripts. Current

Had an amazing time working on this with Danish Pruthi and Mansi ! Catch janki nawale and me at our poster at ACL 2025 in Vienna - would love to chat!

Reading books is perhaps the best training we’ve got to prevent and cure the constantly distracted mind.

Wow! Prof Mitesh Khapra's name is in Times 100/AI in the same list as Fei-Fei Li, Sam Altman, Elon Musk, Jeff Dean and so on. I bless the day I was invited to work alongside him in AI4Bharat! Truly a visionary! time.com/collections/ti… Reply 🐐 to show respect!

No one has changed the course of my life like this man. From day one, his work ethic stood out, if even 1% of that rubs off on me, it’s enough for a lifetime. No better day to say: O Captain! My Captain! Proud to be your student, today and always!

A moment of global pride for India in the field of AI. Prof. Mitesh M. Khapra, Co-founder at the Nilekani Centre at AI4Bharat, WSAI, IIT Madras, has been featured among '2025 TIME100 AI List of the World’s Most Influential People in Artificial Intelligence' AI4Bharat is

The scale of IndicVoices is powered by people. Nearly 1,900 coordinators, mobilizers, and experts worked alongside 16,000+ participants to capture speech in 22 Indian languages. Their role wasn’t limited to logistics. They built trust, helped participants navigate apps, and

🚀Introducing Voice + Chat AI Agents on WhatsApp with Sarvam Samvaad Now create rich customer interactions within one seamless thread that includes voice calling + real-time chat. This is powered by: ▪️11 Indian languages across text, voice and widgets ▪️Context-aware,

Prof. Mitesh M. Khapra, Co-founder at the Nilekani Centre at AI4Bharat, WSAI, IIT Madras, feature on '2025 TIME100 AI List of the World’s Most Influential People in Artificial Intelligence' highlights his groundbreaking contributions to narrowing the AI divide for Indian

Professor Pushpak Bhattacharyya passed away this morning. This world has lost a great human being and a researcher. May he rest in peace.

Experience Sarvam at Global Fintech Fest 2025. Come meet our team and see how we’re playing a part in India’s financial sector’s AI journey. We’re rolling out new product launches, showcasing cutting-edge innovations, and unveiling Sarvam’s latest capabilities for the BFSI