Profile banner
Saharsha Navani

Saharsha Navani

AI/ML researcher and entrepreneur at the forefront of innovation

Chester Springs, PA
rocketreach
Joined December 2024
Favicon imagelinkedin.com
Favicon imagegithub.com
Favicon imageopenreview.net
Favicon imageen.unesco.org
Favicon imagepapers.ssrn.com
Favicon imagepapers.ssrn.com
Favicon imagearxiv.org
Favicon imageresearchgate.net

Network

4.2K connections
🔔
University Pennsylvania Network
💻
IBM AI Researchers
🧠
Google DeepMind Scientists
👑
AI Research Leaders
🚀
AI Startup Founders
📈
Scale AI Professionals
🌱
...

Summary

Saharsha Navani is a highly accomplished AI/ML researcher and engineer with a strong focus on probabilistic learning, AI governance, and mechanistic interpretability. His research includes contributions to works on emergent world representations in VLA models, the generalizability of linear probing, and surveys on model-based reinforcement learning. He has been affiliated with prominent research institutions such as the MIT-IBM Watson AI Lab, Google DeepMind (AlphaFold team), LSE.AI, and ERA Cambridge, indicating a deep engagement with cutting-edge AI advancements and a commitment to understanding and shaping the future of AI. openreview+7
A natural entrepreneur, Saharsha co-founded inQ Technologies, an AI startup that aimed to revolutionize research productivity through holistic OS-scoped RPA AI. The company achieved a $10 million valuation and was backed by notable institutions like UPenn, Stanford, Dell, and MIT, eventually progressing towards an acquisition by an M&A group. His current role as Co-Founder and CTO at Lexor, along with his experience as a Venture Partner at V11 and Contrary, further highlights his entrepreneurial drive and involvement in the venture capital ecosystem. boldjourney+3
Saharsha is deeply involved in AI policy and ethics, serving as an ECOSOC Youth Delegate to the United Nations Department of Economic and Social Affairs and as a core member of the UN IGF Dynamic Teen Coalition. In these roles, he collaborates with global leaders, including Vinton Cerf, on critical discussions surrounding global AI legislation. His philosophical interests, including an early exploration of Stoicism and a current pursuit of General Philosophy alongside Computer Science at Princeton, underscore his commitment to the ethical implications and societal impact of AI. boldjourney+2
Recognized as one of Greater Philadelphia's top young innovators, Saharsha was featured in the 'Inno Under 25' class of 2025 by the Philadelphia Business Journal. This recognition, alongside his Lifetime Fellowship at The Germination Project, showcases his early impact and potential in both technological and societal spheres. bizjournals+1

Work

Education

Writing

Emergent World Representations in OpenVLA

September 1, 2025

Research on Vision Language Action (VLA) models and their emergent world representations.

Favicon imagearxiv.org

Analyzing the Generalizability of Linear Probing and...

September 1, 2025

Research on the generalizability of linear probing in AI models.

Favicon imageopenreview.net

Comment on 2025 National Artificial Intelligence (AI) Research and Development (R&D) Strategic Plan

June 1, 2025

A comprehensive analysis of the NSF's 2025 AI R&D Strategic Plan, examining its strengths and potential weaknesses, and suggesting refinements for equitable, accountable, and innovative execution.

Favicon imagepapers.ssrn.com

Comment on the NIST AI Standards Zero Drafts Pilot Project

June 1, 2025

A comment on the NIST AI Standards Zero Drafts pilot project, emphasizing the importance of balancing innovation with standardization.

Favicon imagepapers.ssrn.com

Emergent Linear Representations in World Models of Self-Supervised Sequence Models

December 1, 2023

A paper exploring emergent linear representations within world models of self-supervised sequence models.

Favicon imageresearchgate.net

Model-based Reinforcement Learning: A Survey

December 1, 2022

A survey of model-based reinforcement learning techniques.

Favicon imageresearchgate.net

Hobbies

Saharsha enjoys playing chess. github