About me

I’m Arkajyoti (AJ) Bhattacharjee, a Ph.D. candidate in Statistics at The Ohio State University, advised by Prof. Arnab Auddy.

My research develops statistical inference for unsupervised learning and generative models, with a focus on privacy, comparison, and memorization. Much of my work centers on density modes, the local peaks of a probability density, which represent characteristic patterns in data, such as the distinct gene-activity profiles of different cancer types. I study how to estimate all of these modes while formally protecting the privacy of every individual record, and how to test whether the modes of two populations are in the same locations even when the share of observations near each mode differs. Privacy concerns also arise in generative models: a model trained on patient scans can produce a synthetic image that reproduces a real patient’s scan. I develop tests that distinguish this memorization from the natural similarity between independent individuals, along with theory that explains when memorization occurs. For each of these problems, I prove theoretical guarantees and evaluate the methods on real data.

I grew up in Kolkata, India, the city of sweets, and received my B.S. in Statistics from Presidency University, Kolkata. I then went on to IIT Kanpur for my M.S. in Statistics, where working with Prof. Dootika Vats on adaptive Markov chain Monte Carlo gave me my first experience of research. During my master’s, I also spent a summer putting statistics into practice. Through Google Summer of Code, I worked on improving R’s nonlinear least-squares tools, which led to a paper in The R Journal. At Accenture, I built an AutoML tool in Python as a data science analyst intern. It was my research at IIT Kanpur, though, that led me to pursue a Ph.D., and in 2022 I came to Ohio State. Since then, I have also completed a graduate minor in Computer Science, and in Summer 2026 I returned to industry as a Summer Associate at JPMorgan Chase, building machine-learning models for credit-risk decisions.

Outside of research, I spend most of my free time strength training.

News

• May 2027 — Received the Ransom and Marian Whitney Award for Research, OSU Department of Statistics.
• Oct 2026 — Won the Graduate category of the ASA Columbus Chapter Scientific Communication Video Contest (video).
• Aug 2026 — One of 6 finalists in the ASA Nonparametric Statistics Section Student Paper Competition; presented at JSM 2026.
• Aug 2026 — Elected president of the (Bio)Statistics Graduate Student Association.
• Summer 2026 — Summer Associate at JPMorgan Chase; passed the Ph.D. candidacy exam.
• May 2026 — Won the People’s Vote Poster Award at the ASA Midwest Regional Conference.
• Apr 2026 — Won 1st Place Poster at the Annual Joint Biostatistics Symposium.

Research interests

• Differential privacy
• Nonparametric and modal inference
• Two-sample testing
• Memorization in generative models
• Tree-based predictive modeling