About

I am a third-year PhD student in Computer Science at MIT CSAIL. I am advised by Bonnie Berger. I also collaborate closely with Emma Pierson. I am broadly interested in machine learning and decision-making, especially in biomedical and healthcare problems. I am particularly interested in how we can make decisions in high-stakes healthcare settings in the real world, where datasets are noisy, sparsely labeled, and replete with biases. Particular problem areas of interest include:

• Sequential decision-making: how can we build models for chained sequences of human decisions when some outcomes are censored as a result of those decisions? For instance, one cannot observe a cancer diagnosis unless the individual passes through several rounds of screening.
• Uncertainty quantification and calibration: how can we quantify uncertainty over predictions in an efficient and correct way? Are there tradeoffs between calibration, optimal (downstream) decision-making, accuracy, etc and how can we quantify them?
• Evaluations: how can we evaluate the best model for a task given constraints on resources (e.g. few labels)?

Prior to MIT, I was an undergraduate at Harvard, where I received an AB in Computer Science and Statistics. Towards the beginning of my PhD and end of undergrad, I worked on biomedical data privacy. Prior to that, I worked on statistical genetics and computational physics. I am fortunate to be supported by the Hertz Foundation Fellowship and NSF Fellowship.

Outside of work, I enjoy hiking, biking, and rooting for New England sports teams. Hit me up if you’re in Boston and want to join on weekend bike rides. I am always happy to chat, feel free to reach out via email at ssadhuka (at) mit (dot) edu.

Publications (* denotes equal contribution)

(Working papers and preprints listed in CV)

• Privacy-Enhancing Technologies in Biomedical Data Science [Link]
  H. Cho, D. Froelicher*, N. Dokmai*, A. Nandi*, S. Sadhuka*, M. Hong*, B. Berger
  Annual Reviews of Biomedical Data Science 2024

• Accurate Evaluation of Transcriptomic Re-identification Risks Using Discriminative Sequence Models [Link]
  Shuvom Sadhuka, Daniel Fridman, Bonnie Berger, Hyunghoon Cho
  Genome Research 2023 and oral presentation at Research in Computational Molecular Biology Conference (RECOMB) 2023

• Topological Phononic Logic [Link]
  Harry Pirie, Shuvom Sadhuka, Jennifer Wang, Radu Andrei, Jennifer Hoffman
  Physical Review Letters 2022

• Leveraging Supervised Learning for Functionally Informed Fine-Mapping of cis-eQTLs Identifies and Additional 20,913 Putative Causal eQTLs [Link]
  Qingbo Wang, David Kelley, Jacob Ulrisch, Masahiro Kanai, Shuvom Sadhuka, Ran Cui, Carlos Albors, Nathan Cheng, Yukinori Okada, Biobank Japan Project, François Aguet, Kristin Ardlie, Daniel MacArthur, Hilary Finucane
  Nature Communications 2021

Other (Random) Writing

• Fellowship Advice [Link]
  Some advice on applying for graduate fellowships!

• Overcoming the False Tradeoff in Genomics: Privacy and Collaboration [Link]
  Honorable Mention at MIT Envisioning the Future of Computing Essay Prize 2023

• Does the New NBA Draft Lottery Systems Really Discourage Tanking? [Link]
  Harvard Sports Analysis Collective 2019