Sankalp Garg

I am the co-founder of Optexity, where I am training AI to use computers just like humans. Previously, I worked at Apple on the Siri Info Intelligence Team. I hold an M.Sc in Machine Learning from Carnegie Mellon University where I was advised by Prof. Zico Kolter and Prof. Aditi Raghunathan.

Prior to CMU, I was a Research Fellow at Microsoft Research, India advised by Prateek Jain and Harsha Vardhan Simhadri. I worked on EdgeML, developing ML algorithms for severely resource-constrained devices.

Before this, I completed a B.Tech in Electrical Engineering from IIT Delhi. I was advised by Prof. Mausam and worked on Reinforcement Learning and Transfer Learning for Planning problems.

Scholar / LinkedIn / X

Research

My research interests lie in the intersection of Machine Learning and Reinforcement Learning. I am particularly interested in the application of RL to real-world problems, such as robotics, autonomous systems, and decision-making under uncertainty. I have worked on Foundation Models, Large Language Models (LLM), and CLIP. I am also interested in the development of algorithms that can learn from limited data and generalize to new tasks.

	Finetune like you pretrain: Improved finetuning of zero-shot vision models Sachin Goyal, Ananya Kumar, Sankalp Garg, Zico Kolter, Aditi Raghunathan Computer Vision and Pattern Recognition Conference (CVPR), 2023 arXiv / Talk / Github
	Symbolic Network: Generalized Neural Policies for Relational MDPS Sankalp Garg, Aniket Bajpai, Mausam International Conference on Machine Learning (ICML), 2020 arXiv / Talk / Github
	Attribute Prediction via Joint Modeling of Multi-Relational Structure Evolution Sankalp Garg, Navodita Sharma, Woojeong Jin, Xiang Ren International Joint Conference on Artificial Intelligence (IJCAI), 2020 arXiv / Github
	Size Independent Neural Transfer for RDDL Planning Sankalp Garg, Aniket Bajpai, Mausam International Conference on Automated Planning and Scheduling (ICAPS), 2019 arXiv / Github
	Transfer of Deep Reactive Policies for MDP Planning Aniket Bajpai, Sankalp Garg, Mausam Conference on Neural Information Processing Systems (NeurIPS), 2018 arXiv / Talk / Github

Employment

Optexity August 2025 - Present	Co-founder Building agents that train AI to use computers just like humans.
Apple Jan 2023 - August 2025	Siri Info Intelligence Team Language understanding and search ranking for Siri, Spotlight, and Safari.
Amazon AI May 2022 - Aug 2022	RL-based Rightsizing of Redshift Databases Developed RL methods for database sizing and built a simulator for training RL models. Advisor: Dr. Murali Narayanaswamy
Quadeye July 2020 - Aug 2021	Quantitative Strategist Developed and managed trading strategies in India and Brazil.
Microsoft Research India Jan 2020 - June 2020	Keyword Spotting on Tiny Devices Built phoneme detection and keyword classifiers for low-memory devices. Advisors: Prateek Jain and Harsha Vardhan Simhadri
University of Southern California Summer 2019	Temporal Knowledge Graphs Designed a framework for time-series prediction using temporal graph embeddings. Advisor: Xiang Ren
National University of Singapore Summer 2018	Explainable AI for Food Recommendations Developed a system for explainable food recommendations and conversational Q&A. Advisor: Prof. Brian Y. Lim

Education

Carnegie Mellon University Aug 2021 - Dec 2022	M.Sc in Machine Learning Machine Learning Department, CGPA: 4.08 Advisors: Prof. Zico Kolter and Prof. Aditi Raghunathan
Indian Institute of Technology Delhi July 2016 - July 2020	B.Tech in Electrical Engineering CGPA: 9.36/10, Department Rank 5 Advisor: Prof. Mausam

Miscellaneous

Conference Reviewer: ICML 2024, ICLR 2024, NeurIPS 2023, ICAPS 2023, ICAPS 2022

Template inspired by: Jon Barron