
I'm building the future of interaction with AI Voices at Kalpa Labs solving hard research problems in modelling voice for long context, ultra-low latency conversational agents & models that can do complex audio editing tasks in one shot.
Previously, I was a Senior Software Engineer at Google, having worked there for 4 years. For the first 2 years, I worked on improving ad retrieval strategies for LocalAds adding $XXXM in annual revenue. For the last two years, I worked at Google Assistant & Home Intelligence where I worked on improving the quality of models & making these models efficient to serve for tens of billions of queries per month with limited TPUs.
You can email me at pshishodia@kalpalabs.ai
Published Papers
Scaling Laws and Efficient Inference for Ternary Language Models
Tejas Vaidhya, Ayush Kaushal, Vineet Jain, Francis Couture-Harpin, Prashant Shishodia, Majid Behbahani, Irina Rish, Yuriy Nevmyvaka
ICML Workshop (Sparsity in LLMs), ACL • 2024
Previous Work
Husn: Realtime Vibe-based search on millions of products
Discover your personal style with Husn's vibe-based search across millions of fashion products.




