San Francisco, CA
Email: abhay.venkatesh[at]gmail[dot]com
LinkedIn ·
X ·
Google Scholar
Abhay is an AI engineer. He was an early ML engineer at Anduril, working on autonomy and later on frontier models for national security. Before that he did machine learning research.
Abhay co-founded Anduril’s Frontier AI team, recruited the founding team, and built AI infrastructure supporting work across the company, including training infrastructure used to post-train models for national security applications.
Related programs: Thunderforge, Anduril’s OpenAI partnership, and CDAO’s Edge Data Mesh for generative AI.
Abhay worked directly with CEO Brian Schimpf to co-invent and patent foundational systems for Lattice for Mission Autonomy, which enables teams of autonomous systems to carry out complex missions under human supervision. Lattice for Mission Autonomy has since been flight-tested on Fury (YFQ-44A), Anduril’s aircraft for the U.S. Air Force’s Collaborative Combat Aircraft program.
Abhay is a key architect of Lattice Command & Control (C2) and a top code contributor to Anduril’s APIs. He architected the tasking APIs, which let operators and software create, route, and track tasks across assets and teams of assets.
RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement
Will LeVine, Brendan Evers, Sam Saltwick, Abhay Venkatesh
arXiv preprint 2026
paper
Generating Accurate Pseudo-labels in Semi-Supervised Learning and Avoiding Overconfident Predictions via Hermite Polynomial Activations
Vishnu Suresh Lokhande, Songwong Tasneeyapant, Abhay Venkatesh, Sathya N. Ravi, Vikas Singh
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2020
paper |
code
Optimizing Nondecomposable Data Dependent Regularizers via Lagrangian Reparameterization Offers Significant Performance and Efficiency Gains
Sathya N. Ravi, Abhay Venkatesh, Glenn M. Fung, Vikas Singh
AAAI Conference on Artificial Intelligence (AAAI) 2020
paper
Generating Accurate Pseudo-labels via Hermite Polynomials for SSL Confidently
Vishnu Suresh Lokhande, Sathya N. Ravi, Songwong Tasneeyapant, Abhay Venkatesh, Vikas Singh
Midwest Machine Learning Symposium (MMLS) 2019, Best Student Paper
Nondecomposable Data Dependent Regularizers Offer Significant Performance Gains
Sathya N. Ravi, Abhay Venkatesh, Glenn M. Fung, Vikas Singh
Conference on the Mathematical Theory of Deep Neural Networks (DeepMath) 2019
The Case for Unifying Data Loading in Machine Learning Clusters
Aarati Kakaraparthy, Abhay Venkatesh, Amar Phanishayee, Shivaram Venkataraman
USENIX Workshop on Hot Topics in Cloud Computing (HotCloud) 2019
paper
Learning View Invariant Semantic Segmentation for UAV Video Sequences
Abhay Venkatesh, Glenn Fung, Vikas Singh
SDM 2018 Workshop on Artificial Intelligence in Insurance
paper
Abhay invests in AI and frontier technology. Since 2022, he has invested across 63 companies, including Mercor, Factory AI, Saronic, OpenAI, and Anthropic. He has participated in scout programs at Greylock, Sequoia, Index Ventures, and Bain Capital Ventures.
He also founded Defense Angels, the ex-Anduril angel syndicate.
Abhay founded:
Abhay Venkatesh, Brian W. Schimpf, Elston Tochip, Jared Newman, Joshua John Bennett, Nikhil Reddy, Samuel R. Elie