Neev Parikh

Neev Parikh

Software Engineer

Stripe, Inc.

Biography

Currently, I’m a software engineer at Stripe, Inc. on the financial data team. I graduated from Brown University with a masters (advised by Professor George Konidaris) in Computer Sciecne, with research interests in reinforcement learning and distributed, multi-modal machine learning systems.

For a resume, please click here. Get in touch via email. Other links: github, google scholar, linkedin.

Interests

  • Large-scale machine learning
  • Multi-modal models
  • Distributed reinforcement learning
  • Artificial Intelligence

Education

  • M.Sc. in Computer Science, 2022 (Concurrent)

    Brown University

  • B.Sc. in Computer Science, 2022 (Concurrent)

    Brown University

Recent Publications

*equal contribution

Learning Markov State Abstractions for Deep Reinforcement Learning

We introduce a novel set of conditions and prove that they are sufficient for learning a Markov abstract state representation. We then describe a practical training procedure that combines inverse model estimation and temporal contrastive learning to learn an abstraction that approximately satisfies these conditions.
Cameron Allen, Neev Parikh, George Konidaris
December 2021 NeurIPS 2021
PDF Video

Deep Radial-Basis Value Functions for Continuous Control

We introduce deep radial-basis value functions (RBVFs): value functions learned using a deep network with a radial-basis function (RBF) output layer.
Kavosh Asadi, Neev Parikh, Ron Parr, George Konidaris, Michael Littman
February 2021 AAAI 2021
PDF

Graph Embedding Priors for Multi-task Deep Reinforcement Learning

We leverage graph-encoded object priors to capture this property and improve the performance of reinforcement learning agents across multiple tasks. We introduce a novel, flexible architecture that utilizes graph convolutional networks (GCNs), which provide a natural method to combine relational information over connected nodes.
Neev Parikh*, Zachary Horvitz*, Naveen Srinivasan*, Aansh Shah, George Konidaris
December 2020 KR2ML Workshop, NeurIPS 2020
PDF Poster Workshop link

Locally Observable Markov Decision Process

We introduce a novel structured formulation, the Locally Observable Markov Decision Process, which assumes that partial observability stems from limited sensor range—objects outside sensor range are unobserved, but become fully observed once they are within sensor range.
Max Merlin, Neev Parikh, Eric Rosen, George Konidaris
June 2020 PAL 2020 Workshop, ICRA
PDF Video Workshop link

Experience

 
 
 
 
 

Software Engineer

Stripe

Jul 2022 – Present San Francisco, CA
  • SWE in the Financial Data team.
 
 
 
 
 

Research Engineering Intern

Common Sense Machines

May 2021 – Aug 2021 Boston, MA
  • Implemented large-scale, auto-regressive Seq2Seq models for working with 3D geometry from images.
  • Worked with Deepspeed to explore scaling options for 500M+ param models to feasibly scale existing sequence-based models.
  • Implemented a graphics algorithm to find surface patches in a 3D wire-frame (Zhang et. al., 2013 )
  • Dockerized AWS pipeline to create cloud-independent dev/production
 
 
 
 
 

Research assistant

Intelligent Robot Lab

Jun 2020 – May 2022 Providence, RI

Research assistant at Intelligent Robot Lab (irl) at Brown University, led by professor George Konidaris.

  • Current projects:

    • New mathematical framework (LOMDPs) for tasks in robot domains
    • Unsupervised representation learning for Atari task suite
    • Graph-based priors for improving multi-task performance
 
 
 
 
 

Teaching Assistant

Department of Computer Science, Brown University

Jan 2020 – May 2020 Providence, RI

Worked as a teaching assistant for Computer Vision (CSCI 1430), taught by Professor James Tompkin.

  • Guided 3 teams in their final projects
  • Revamped Project 5 and helped update 4 other projects for the class
  • Held weekly office hours, biweekly grading of projects and homeworks
 
 
 
 
 

Teaching Assistant

Department of Computer Science, Brown University

Sep 2019 – Dec 2019 Providence, RI

Worked as a teaching assistant for Introduction to Reinforcement Learning (CSCI 2951F), taught by Professor Michael Littman.

  • Extended course library framework simpleRL fixing common bugs and improving example visualizations
  • Designed 3 homeworks for the class
  • Guided 3 teams in replicating papers from NeurIPS 2019 for the NeurIPS 2019 Reproducibility Challenge
 
 
 
 
 

Machine Learning Intern

Myelin Foundry

Jun 2019 – Aug 2019 Bangalore, India

Worked on a cutting-edge, deep-learning based pipeline to augment VFX workflows.

  • Worked to develop cutting-edge, deep-learning based pipeline to augment VFX workflows for a POC product
  • Researched and managed a company-wide, cloud-compute platform, reducing potential monthly costs by 70%
  • Helped transition MLOps to Microsoft Azure
  • Implemented DeepLabv3+ from ECCV 2018 to develop SOTA pipelines for semantic segmentation tasks
  • Achieved 90% in business-aligned metrics with reasonable inference time
 
 
 
 
 

Software Engineering Intern

Scripbox

Jul 2018 – Aug 2019 Bangalore, India
  • Developed integrated data visualization tool in Typescript with ReactJS and Django (Python).
  • Researched RFM analysis to gather business insights using Python.