Adithya S Kolavi

Adithya S Kolavi is among the most promising young AI researchers and developers emerging from India’s open-source AI ecosystem. Based in Bengaluru, Kolavi has built a reputation for working on vision-language models, multilingual AI systems, and agentic architectures designed for real-world deployment. He is currently working at Hugging Face. 

He was previously a Research Fellow at Microsoft Research and also interned with Apple on its Unified Intelligence team, a trajectory that reflects strong recognition from some of the world’s leading technology companies.

Kolavi is also deeply connected to India’s open-source AI movement through CognitiveLab, a Bengaluru-based research collective focused on democratising AI for Indic languages. 

At CognitiveLab, he helped develop Ambari-7B, one of India’s earliest Kannada-English bilingual large language models. He has also contributed to projects including OmniParse, a document-processing library that gained significant traction among developers, and Nayana OCR to improve document understanding for low-resource Indian languages.

Kolavi has published and contributed to projects spanning OCR, multimodal reasoning, synthetic dataset generation, and document-centric AI systems. He also actively works on agentic workflows, model fine-tuning, and LLMOps infrastructure, areas becoming central to enterprise AI adoption.

He is currently a research engineer at Hugging Face.