Summary
Himansu Didwania is a Machine Learning Researcher with nine years of experience specializing in voice AI and Text-to-Speech, currently leading TTS research as a founding member at Sarvam AI in Bengaluru. He has a strong applied-research background building end-to-end voice systems—spanning ASR, intent recognition, entity extraction, and SLU improvements—at AI4Bharat and Skit.ai. His work blends research and product delivery: from defining metrics and deployment practices for voice bots to co-founding a startup that generated talking avatars from text. Comfortable moving models from prototype to production, he also has academic research experience in robotics, generative models, and scene text detection from internships at IIIT Hyderabad and ISI Kolkata. Colleagues would describe him as a practitioner who pairs deep technical curiosity with a bias for operationalizing speech research into user-facing systems.
9 years of coding experience
3 years of employment as a software developer
Bachelor of Technology, Information Technology, Bachelor of Technology, Information Technology at International Institute of Information Technology, Bhubaneswar