Varun Nagaraj Rao

Varun Rao

PhD Candidate in Computer Science

Princeton University

Email: firstnamelastname@princeton.edu

About

I am a computer scientist and social computing researcher with a passion for understanding and mitigating the societal impacts of AI, particularly in the context of labor and technology policy. My research spans across societal impacts of AI, explainable AI, and multimodal machine learning.

Bio

Varun Rao, is a Ph.D. student in Computer Science at Princeton University advised by Prof. Andrés Monroy-Hernández, currently part of the Workers Algorithm Observatory (WAO). His research combines system building and empirical studies to understand and mitigate AI's impact on labor. Current projects include FairFare, examining black-box rideshare algorithms, and OpenDeli, exploring decentralized food delivery platforms. Earlier in his Ph.D., Varun's work exposed discrimination in job ad image selection on platforms like Facebook, revealing how this practice circumvents existing protections.

In his professional experience, Varun worked as an Applied Scientist at Amazon, developing explainable vision-language models for customer protection and seller trust. During his internship at Apple, he enhanced OCR solutions for handwritten text recognition

Varun engages in tech policy outreach, influencing legislation on rideshare transparency. He is passionate about CS Education Research focusing on introductory algorithms courses and developing concept inventories. His research has earned several recognitions, including the COLING 2020 Outstanding Paper Award, MIT/DARPA Graph Challenge Honorable Mention 2017, and the Hamsa Kartik Alumni Award for ranking first in his undergraduate class.

Affiliations at Princeton: HCI Group, Center for Information Technology Policy (CITP), Computer Science

News

Education

Princeton University

Doctor of Philosophy in Computer Science, Aug 2022 - Present

Research: Societal Impacts of AI on Labor

Selected Coursework: HCI (TA), Technology Policy and Law, Privacy, Ethics of Computing (TA), Responsible AI (TA)

Carnegie Mellon University

Master of Science in Electrical and Computer Engineering, Dec 2019

GPA: 3.96/4.0

Coursework: Computer Vision, Machine Learning (PhD), Convex Optimization, Security and Fairness in Deep Learning, Advanced Multimodal Machine Learning, Foundations of Cloud and ML Infrastructure

PES Institute of Technology

Bachelor of Engineering in Computer Science and Engineering, May 2018

GPA: 9.97/10 (Rank 1/132)

Delhi Public School Bangalore North

CBSE Class XII (Science), May 2014

Score: 96.6% (Rank 1)

Vidyashilp Academy

ICSE Class X, May 2012

Score: 98.8% (Rank 1)

Ongoing Projects

Workers Algorithm Observatory (WAO)

Understanding gig worker concerns and building tools to measure and mitigate AI and algorithmic systems harms; Focus on rideshare and delivery platforms. Contributed to FairFare Project.

Academic paper in progress

Project Website

OpenDeli

Designing a decentralized food delivery protocol and a reference implementation consisting of a frontend mobile app, backend and admin interface.

Academic paper in progress

Project Website

Publications

Societal Impacts of AI and Technology

Explainable AI and Multimodal Machine Learning

High Performance Computing

Computer Science Education

Teaching Experience

Professional Experience

Applied Scientist II Intern

Amazon Web Services, AWS Bedrock

May - Aug 2023

  • Showed that captioning, VQA and image classification can be modeled through a unified retrieval augmented encoder-decoder vision-language architecture, with no pretraining and additional trainable parameters; paper under review.
  • Experiments demonstrate the benefits on image captioning (+1 CIDEr on MSCOCO, +4 CIDEr on NoCaps) and VQA (+3 % accuracy on VQA v2) tasks compared to the non-retrieval baselines.

Applied Scientist II

Amazon.com Inc, Multimodal AI - Product Assurance, Risk and Security (PARS)

Feb 2020 - July 2021

  • Explainable multimodal text-in-image and classification models to help protect customers and build seller trust.
  • Invited Talk "Explainable Multimodal TextVQA Models" - Amazon Machine Learning Conference (AMLC) 2020
  • Outstanding Paper Award for "Misspelling Detection from Noisy Product Images" (top 2.5% - 16/644) at COLING 2020

Computer Vision Intern

Apple, Inc - Core Recognition Team

May - Aug 2019

  • Enhanced the accuracy of the OCR system by up to 28% and expanded support for handwritten text.

Software Development Intern

Akamai Technologies - Analytics Database Team

Feb - Apr 2018

  • Designed and implemented a release checklisting framework that reduced the time for checklisting in a sprint by up to 2 days.

Awards

Broader Outreach

Service

Reviewer / Program Committee Member

Invited Talks

Workshops and Summer Schools

Mentorship

I am fortunate to work or have worked closely with many excellent undergrad students at Princeton: