Welcome! Come in :)
Welcome! Come in :)
Who am I?
I’m Satchit Sivakumar, a fifth-year Ph.D. student in Computer Science at Boston University, advised by Mark Bun. My research centers on responsible machine learning, with a focus on the mathematical foundations of data privacy and replicability, and their connections to statistics, machine learning, and public policy. Click here to see my research publications on these topics!
I am honored to have my work supported by the Apple Scholars in AI/ML PhD fellowship. I also was a Machine Learning Research Intern at Apple in 2023 and 2024, where I worked with Vitaly Feldman, Audra McMillan, and Kunal Talwar on differentially private statistics.
I am on the job market for Summer/Fall 2026- looking for Postdoc and Industry research positions.
If you'd like to connect, or chat about opportunities, please reach out! My email is satchit@bu.edu
Interactive Proofs For Distribution Testing With Conditional Oracles
with Ari Biswas, Mark Bun, Clément Canonne.
In Submission to ITCS 2026. Manuscript available upon request.
Improved Accuracy for Private Continual Cardinality Estimation in Fully Dynamic Streams via Matrix Factorization
with Joel Andersson, Palak Jain.
In Submission to SODA 2026. Manuscript available upon request.
Enforcing Demographic Coherence: A Harms Aware Framework for Reasoning about Private Data Release
with Mark Bun, Marco Carmosino, Palak Jain, Gabriel Kaptchuk.
Spotlight Talk at TPDP 2025.
Instance Optimal Private Density Estimation in the Wasserstein Distance
with Vitaly Feldman, Audra McMillan, and Kunal Talwar.
NeurIPS 2024. Spotlight Talk at TPDP 2024.
Counting Distinct Elements in the Turnstile Model with Differential Privacy under Continual Observation
with Palak Jain, Iden Kalemaj, Sofya Raskhodnikova, and Adam Smith.
NeurIPS 2023. Oral Presentation at FORC 2024. TPDP 2023.
Stability is Stable: Connections between Replicability, Privacy, and Adaptive Generalization
with Mark Bun, Marco Gaboardi, Max Hopkins, Russell Impagliazzo, Rex Lei, Toniann Pitassi, and Jess Sorrell.
STOC 2023, Spotlight Talk at TPDP 2023, Oral Presentation at FORC 2023.
The Price of Differential Privacy under Continual Observation
with Palak Jain, Sofya Raskhodnikova, and Adam Smith.
Oral Presentation at ICML 2023 (2.37% of all submissions), Spotlight Talk at TPDP 2022, FORC 2023.
Differentially Private Sampling from Distibutions
with Sofya Raskhodnikova, Adam Smith, and Marika Swanberg.
SICOMP 2025, NeurIPS 2021. Spotlight Talk at TPDP 2021.
Multiclass versus Binary Differentially Private PAC Learning
with Mark Bun and Marco Gaboardi.
NeurIPS 2021. Poster at TPDP 2021.
May 2025: Honored to win the BU CS research excellence award for my PhD work on privacy-preserving data analysis.
February 2025: New paper on arXiv suggesting a new framework for reasoning about the privacy of large data releases! With the goal of tying more intuitively to privacy harms and thereby allowing for easier reasoning about parameters.
January 2025: Visited UC San Diego for a workshop on new definitions/approaches to privacy that moved us closer to practice! Realized how much I'd missed shorts weather in January :)
December 2024: Presented our work on instance optimality at NeurIPS 2024 in Vancouver! From what I can tell, Vancouver has no distinctive food identity.
August 2024: Gave a spotlight talk at TPDP 2024 on instance optimality!
June 2024: New paper on arXiv that defines instance optimality for density estimation tasks and gives near instance optimal algorithms (both privately and nonprivately) for this problem in some settings!
May 2024: I will be once again be a Machine Learning research intern at Apple, Cupertino this summer! Working on private computation of statistics on stratified sampling designs :)
November 2023: Gave a TCS+ talk on my line of work on differentially private continual release!
September 2023: Gave a spotlight talk at TPDP 2023 on replicability!
June 2023: Conference season, baby! Gave talks at FORC 2023 on the continual observation model (with my friend Palak) and on replicability. Also spoke at STOC 2023 on replicability. Had tons of fun and learned a lot at both conferences! Thanks to the organizers and the brilliant people I met.
June 2023: New paper on counting distinct elements in the continual observation model.
May 2023: I will be a Machine Learning research intern at Apple, Cupertino this summer! Reach out if you're in or around the Bay Area and would like to chat.
April 2023: Our paper on the continual release model of differential privacy was accepted for an oral presentation at ICML 2023!
February 2023: Proud to have been chosen as a recipient of the 2023 Apple Scholars in AI/ML PhD Fellowship.
February 2023: Woooo! Our paper on connections between differential privacy, replicability, and other notions of stability will appear in STOC 2023! Here I come, Orlando!
September 2022: I'm a Teaching Fellow for CS332:' Theory of Computation ' this semester, and am fondly remembering my love for DFAs and NFAs as an undergrad.
August 2022: I was selected to attend a summer school on 'New tools for Optimal Mixing of Markov Chains: Spectral Independence and Entropy Decay', at the beautiful UC Santa Barbara campus. The campus has one of my favorite smoothie places, with an unfortunately mediocre name, 'Blenders in the Grass'.
July 2022: I was invited to attend a workshop on 'Differential Privacy and Statistical Data Analysis' at the Fields Institute and gave a talk on 'Differentially Private Sampling from Distributions'! My main thought about Canada: I do not understand the hype surrounding Poutine.
July 2022: I attended and gave an invited talk on 'The Price of Differential Privacy under Continual Observation' with my good friend Palak Jain at TPDP 2022 in Baltimore! Baltimore has the best vegan soul food.