Mehtaverse

Bio

I am a postdoctoral associate at Boston University and MIT developing mechanisms for reducing strategic deception in information environments where humans interact with AI agents. I explore how credibility signals shape human behavior on digital platforms. My research explores the social, technological, and economic drivers of digital trust.

Here’s my job market paper on “Market Design Interventions for Safer Agentic AI” and a first-author working paper on time-series causal inference for social networks.

Applied Research

I founded and co-lead a tech nonprofit, SimPPL, rebuilding trust in online information, collaborating with international newsrooms, think tanks, and UN agencies in several countries. We raised $2.5 million in competitive awards with collaborators, working on two projects:

1. Arbiter is a platform to trace digital discourse across social media platforms, especially for the early detection of digital harms. We empower researchers to trace claims across digital platforms including Bluesky, YouTube, Reddit, X, and others, improving transparency in the online information environment.
2. Sakhi is an award-winning, MIT-incubated tech venture delivering reproductive health literacy in India and Bangladesh, now working with Cohere Labs on reproductive health evaluations of large language models.

While Arbiter improves digital transparency, Sakhi explores trust-building mechanisms in health information.

Professional Background

In industry, I worked at Twitter, Adobe, Slack, with Meta at Oxford, startups, and CERN on AI products, algorithms, and applications. I have been an Atlantic Dialogues Emerging Leader, Google Research Innovator, CTS Belfer Fellow, and won several grants from Google, Mozilla, Wikimedia, Ford Foundation, Omidyar Network, and others to support my research involving SimPPL.

Dissertation: I received my Ph.D. (2019-23) from NYU Data Science, estimating the causal effects of interventions to improve online safety. My research was one of the first to show platforms that the interventions they deploy can have meaningful effects off-platform even if they seem ineffective within their platform, and resulted in evantual internship offers at Meta, Twitter, Mirosoft, and Google.

Advising: I served as a Community Advisory Board member at the Integrity Institute (read the 2023 Annual Impact), Responsible Tech Affiliate at All Tech is Human, and sit on 3 Boards of Studies at D.J. Sanghvi College of Engineering. I’ve mentored about 130 undergraduates and 4 high-school students. 4 of my mentees are now pursuing their Ph.Ds in Computer Science while several are graduate students, and others work at FAANG companies and startups.

Teaching

I have run several bootcamp-style programs spanning a week to a year, training several hundred students in computer science, programming, machine learning (ML), artificial intelligence (AI), and most recently on AI Product Development. These programs won SimPPL a unique double award from Google exploreCSR (twice) and Mozilla Responsible Computing Challenge. Teaching evaluation via student testimonials are here. My mentees have published 15+ papers at ICML, IC2S2, ICWSM, NeurIPS, AAAI, Stanford Trust and Safety, Dg.O, and other workshops/conferences.

Art: I don’t talk much about this but I kinda went to 14 years of art school, some of my work is here.