Hi! I'm

Abhinav Raghunathan

I'm a data scientist and engineer focused on responsible AI.
I research the ethical AI startup ecosystem for the Ethical AI Database.

BS Computational Engineering + Mathematics @ UT Austin '21
MSE Data Science @ University of Pennsylvania '22


Fintech Engineering Intern

June 2019 - August 2019

At IHS Markit, I added over $3M to the company's Wall Street Office (WSO) product revenue by developing an automated loan tracking / reconciliation tool for internal purposes. The tool was built to handle several complicated client-specific enhancements and is estimated to have reduced 50+ hours of manual labor for each reconciliation cycle.

Summer Analyst

June 2020 - August 2020

At Point72 Asset Management, I conducted in-depth stock research and offered recommendations to senior analysts and portoflio managers. The analyses were a mixture of qualitative (DCF, comparables, three-statement models) and quantitative (hypothesis testing) components. I also leveraged large datasets from third-party vendors (e.g., AppAnnie) to motivate my investment theses.

Data Science Intern, Data Scientist

February 2022 - May 2022 (joining full-time)

At Vanguard IMFS (Investment Management Fintech Services), I built an API (Python) for several common use cases that allowed the team to access and implement Refinitiv financial data at a much faster pace. I also carried out several statistical tests and backtests (Python, time series, online learning) on order flow imbalances to identify potential investment signals.

Data Science Contractor

February 2022 - April 2022

Though I only worked with Starry from the contractor position, I audited and improved one of their random forests models (R, Snowflake) that was to be placed in production to identify high-reward marketing targets. The resulting model (after I had improved it and checked relevancy) could predict raw subscriber increases with an error rate of \(\pm 1\) subscriber >80% of the time.

Data Science Intern

June 2022 - December 2022

At FairPlay AI, I worked primarily on the Home Mortgage Data Act (HMDA) datasets in order to characterize the state of mortgage fairness in the country. I examined intersectional mortgage fairness as well as geographic fairness differences in lending markets and generated insights that would later be converted into thought leadership pieces for the company (the culmination of my work can be found here, though press coverage of this report can be found on Forbes, USA Today (paywall), and Yahoo Finance.). Later, I worked on creating a baseline framework with custom metrics for marketing fairness (i.e., how fairly is a lending institution marketing their services across represented and underrepresented populations?). This analysis is the first of its kind.


Ethical AI Database (EAIDB)

The only publicly available, vetted database of 200+ AI startups providing ethical services. EAIDB has received recognition from Open Data Science and the Montreal AI Ethics Institute. EAIDB's reports have received over 200 downloads.

Sports vs. Happiness

A study investigating the relationship between the nature of a country's sports teams and that country's overall happiness.


An image recognition algorithm platform meant to process conservationists' photos and identify characteristics such as type of animal, number of animals, etc. and place bounding boxes around the subjects.

Designing an All-Terrain Rover

A research project conducted with IEEE and Winlab at Rutgers University. The goal was to devise a rover with a unique locomotion method that utilized a combination of linear actuators and wheels to traverse tough terrain.


I write broadly about data science and machine learning but tend to focus on topics involving bias / fairness. My research has received acclaim across social media and various other blogs and channels.

TEDx Talk

I gave a talk at UT Austin's TEDx conference in 2021 on the dangers of algorithmic bias and some possible ways to mitigate them. See the slide deck here.

© 2022 Abhinav Raghunathan (abhirags@seas.upenn.edu). Made with Bootstrap.