Call for Participation
Utkarshini
The Vedas have waited three thousand years to be read this carefully. Utkarshini is the first AI model trained deeply and specifically on the Vedas and Upanishads. We are building the scholar-reviewed parallel corpus that makes it possible — and we need people who care about getting it right.
108
Upanishads to be represented in the parallel corpus
1B+
People whose cultural heritage has no reliable AI representation
About this project
We are looking for a Master's Sanskrit Scholar to help build the first AI trained on the Vedas and Upanishads.
This role is unlike any other:
2-month paid on-site internship · Ajmer, Rajasthan
Validate Sanskrit alignments & review AI training data
Utkarshini, from Sanskrit, signifying excellence, brilliance, and one who lifts up, is Bhaskar Lab's flagship Indic LLM project. It is two things at once: a fine-tuned 8-billion-parameter AI model capable of answering complex philosophical questions, citing specific verses, and working across Advaita, Dvaita, and Vishishtadvaita traditions; and the structured parallel corpus beneath it - every verse aligned across Devanagari original, transliteration, word-by-word meaning, and multiple commentarial traditions.
Automated tools cannot assess theological accuracy. The pipeline runs in four phases - each dependent on the last, and each requiring human expertise at its core.
Learn more about the project here.
Get Involved
Ready to contribute to
Utkarshini ?
- ✓This project is bigger than one lab. Building the first structured, machine-readable parallel corpus of Vedic and Upanishadic literature is not a task for a small team alone. The corpus will outlast any single version of the model, and every person who contributes to getting it right is part of what makes it last.
- ✓First-mover significance: No one has built what Utkarshini describes. Zero competition, zero prior art. The foundational contribution is available now.
- ✓Lasting credit: Scholar contributors receive first-author credit on a pioneering AI corpus — publishable at computational linguistics and digital humanities conferences.
- ✓Frontier work, meaningful context: Work at the intersection of deep textual scholarship and frontier AI development, from Ajmer — North India's richest creative crossroads.
Open until 30 Jun 2026
Project Organiser
Adaptiv Admin
@admin
Building the future of AI products at Adaptiv.Me.
More Campaigns
Ajmer Memory Project - Hindi Edition
एक शोध पहल जो अजमेर और उसके आसपास के क्षेत्र के ऐतिहासिक स्रोतों, व्यक्तिगत अभिलेखों और संस्थागत दस्तावेज़ों को मानचित्रित कर रही है — जून में आयोजित होने वाली एक सार्वजनिक प्रदर्शनी की तैयारी में।
Ajmer Memory Project
An initiative to trace, document and preserve the rapidly fading history of the city of Ajmer. This initiative plans to use historical sources, personal archives, and institutional records of Ajmer and its surrounding regions for a free public exhibition in July 2026.
Participation open
Closes 30 Jun
