Mehtaverse

Bio

I am a Ph.D. candidate at NYU Data Science and founder at SimPPL, primarily at CSMAP, collaborating with folks at Oxford TVG and the Integrity Institute. My research deals with limiting disinformation on social networks using tools from machine learning and causal inference:

1. Estimating the Causal Effects of Policy Interventions
2. Auditing the Impact of Content Recommendation Algorithms
3. Building Civic Integrity Tools for Practitioners and Policymakers

I’ve interned with Twitter Civic Integrity, Adobe Research, and worked on machine learning for particle physics at CERN.

I founded SimPPL (read: ‘sim people’), a collective that builds better civic integrity tools such as Parrot Report. We’ve worked with The Sunday Times, Ippen Media, Deutsche Welle, the Yale Daily News, the Vermont Digger, and others on:

1. Supporting newsrooms to better understand their audiences on social networks.
2. Study the spread of unreliable news and identify influence operations online.

SimPPL has received support from the Wikimedia Foundation, Google, AWS, the NYC Media Lab, AI2Amplify (German Federal funding), and the NYU Tech Venture Workshop. It is materially supported by the One Fact Foundation and the AI4ABM Foundation at Oxford.

I mentor students and conduct collaborative research with Unicode Research. This team was born out of the parent org. I co-founded in 2017 to teach open-source development to undergrads, called DJ Unicode.

2023

• Co-authored the Integrity Institute’s:
  • Blog post on GenAI.
  • Transparency Best Practices Deck.
  • Response to Oversight Board on the matter of Meta’s keyword-based enforcement.
• Got a mention on the Google Cloud blog!
• Won the AI2Amplify Fellowship as well as their Hackathon!
• Accepted into the Integrity Institute.
• Received a Google Research Innovator Award.
• Received WikiCred (Wikimedia) funding for expanding Parrot Report with a featured blog post up here
• Hosting multiple talks, workshops for Deutsche Welle Akademie staff and partners (read more)

2022

• Our workshop on AI for Agent-based Models accepted at ICLR 2023!
• Training undergrads from multilingual countries like India to conduct social network analysis via Unicode Research Seminars
• Co-organizing the NeurIPS 2022 Workshop on Broadening Research Collaborations in ML.
• Accepted a position on the Board of Studies at my alma mater, D.J. Sanghvi College of Engineering!
• Invited to present my research on ‘Tools to Limit Disinformation on Social Networks’ at:
  • MIT Media Lab
  • Misinfo. Village @ DEFCON 2022
  • Truth and Trust Online
  • Graphika
• Presented this work on studying the causal effects of Twitter’s interventions on Donald Trump’s tweets at the first Stanford Trust and Safety Research Conference.
• This paper with Oxford and CSMAP studying the effects of influence operations on content ranking and recommendations accepted for an Oral (top 3/26 papers) at the AI for Agent-based Models Workshop at ICML 2022.
• Interning at Twitter’s Civic Integrity team working on the early detection of misinformation in tweets.

2021

• SimPPL selected for NYC Media Lab’s AI and Local News Challenge!
• Early work on SimPPL: Simulating Social Network Interventions with Probabilistic Programs was invited for a presentation to:
  • Twitter’s Cortex Org.
  • Facebook’s Probability Org.
• I received an offer to be a Data Scientist Intern with Microsoft’s Azure ML Team
• Compartmental Models for COVID-19 and Control via Policy Interventions accepted at PROBPROG 2021, MIT.
• “Open Domain Trending Hashtag Recommendation for Videos” with Adobe Research was accepted at the International Symposium for Multimedia, 2021.

2019 - 2020

• I’m a Data Science Research Intern at Adobe Research from May - Dec 2020 at the BEL Lab.
• I’m grateful to have received the IRIS-HEP Fellowship, 2019
• I launched Unicode Research
• For published articles on technology look up my contributions to the Open Source for You (OSFY) Magazine