Vatsal Shah
AI & data leader focused on federated learning and ML
United States
Joined February 2026
Network
6.1K connectionsBJASRK
AMTCJS
BLSBBB
YIPSJM
EHAWMY
JSVZ
BBNM
Summary
Expertise in federated and privacy-preserving ML: Vatsal is focused on decentralized training and federated approaches to reduce data movement and improve privacy (work at FLock.io, writing on membership inference, and DataBloom-related federated data processing initiatives). huggingface+2
Startup operator and technical leader: He has co-founded and served as CTO at EasyQuery and led revenue and growth functions at DataBloom, demonstrating both technical leadership and go-to-market experience in early-stage companies. pangea+1
Active contributor to open-source ML tooling and community projects: Vatsal maintains public code repositories and participates in Hugging Face / FLock-related datasets and models, reflecting hands-on ML engineering and community engagement. github+2
Cross-disciplinary background blending engineering and research: His academic background in mechanical engineering, a visiting research appointment in robotics at the University of Edinburgh, and later ML-focused study (Plaksha/SCET-UC Berkeley) inform work across robotics, ML, and data systems. pangea+1
Work
Education
Projects
Writing
AIArena: A Blockchain-Based Decentralized AI Training Platform
December 1, 2024An academic paper describing AIArena, a decentralized AI training platform enabling on-device optimization and decentralized coordination (arXiv listing includes Vatsal Shah as a co-author).
Membership Inference Attacks in Generative AI: Privacy Risks and Solving It Through Federated Learning
Blog post discussing membership inference attacks against generative models, the privacy risks they pose, and federated learning as a mitigation approach.
Data Strategies in the Wake of AI (white paper / blog)
White paper / blog discussing federated execution, decentralized data strategies, and enterprise data governance in the context of AI.