Aashay Mehta

Aashay Mehta

Principal Data Scientist

Capital One

About Me

Hi, I’m Aashay! I enjoy working on problems with real-world impact. I tend to start by defining the goal and then work backward—methods are tools, not the objective. Given the practical constraints of time and brainpower, I’m most effective when leveraging techniques from computer science, machine learning, and adjacent fields.

I’m currently a data scientist at Capital One, where I work on and with large language models. Before that, I spent three wonderful semesters studying machine learning at Carnegie Mellon University. For more details, feel free to check out my resume.

Outside of work, I enjoy playing chess and racquet sports, attending live concerts and Broadway shows, visiting museums, solving puzzles, exploring the great outdoors, and trying new desserts.

Interests
  • Machine Learning
  • Statistics
  • Software Engineering
Education
  • MS in Machine Learning, 2023

    Carnegie Mellon University

  • BE in Computer Science, 2020

    Birla Institute of Technology & Science (BITS), Pilani

Experience

 
 
 
 
 
Principal Data Scientist
Capital One
Jun 2024 – Present New York, NY
 
 
 
 
 
Graduate Research Assistant
Carnegie Mellon University
Sep 2022 – Dec 2023 Pittsburgh, PA
 
 
 
 
 
Member of Technical Staff
Edgeverve Systems Ltd.
Nov 2020 – May 2022 Bengaluru, India
 
 
 
 
 
Bachelor’s Student Intern
Max Planck Institute for Intelligent Systems
Jan 2020 – Aug 2020 Tübingen, Germany
 
 
 
 
 
Mitacs Globalink Research Intern
University of Alberta
May 2019 – Jul 2019 Edmonton, AB, Canada
 
 
 
 
 
Undergraduate Research Assistant
BITS Pilani
Aug 2018 – May 2019 Pilani, India
 
 
 
 
 
Analytics Intern
Happiest Minds Technologies
May 2018 – Jul 2018 Bengaluru, India

Publications

(2025). On the Query Complexity of Verifier-Assisted Language Generation. ICML.

PDF

(2024). Promises and Pitfalls of Generative Masked Language Modeling: Theoretical Framework and Practical Guidelines. ICML.

PDF

(2023). Data Cross-Segmentation for Improved Generalization in Reinforcement Learning Based Algorithmic Trading. Journal of Technical Analysis.

PDF

(2022). Leveraging machine learning to automatically derive robust decision strategies from imperfect models of the real world. Computational Brain & Behavior.

PDF Project DOI

(2020). Leveraging Machine Learning to Automatically Derive Robust Planning Strategies from Biased Models of the Environment. CogSci 2020.

PDF Project