Davin Choo
Postdoctoral Fellow @ Harvard SEAS

My research centers on developing principled AI and algorithmic solutions for use-inspired, impact-driven problems, organized around three synergistic thrusts that advance both foundations and real-world impact.

Overview of research flow

  • RT1 (Foundational AI/ML): Here, I mainly tackle problems in statistical machine learning and causal inference. I also have some works on clustering, SAT solving, and improving computational subtasks in modern ML pipelines.
  • RT2 (Algorithms with imperfect advice): Here, I design methods that leverage imperfect, instance-specific information to overcome generic worst-case hardness barriers. As part of my PhD, I developed the Test-and-Act framework for designing such learning-augmented algorithms.
  • RT3 (AI+X): Here, I work with domain experts beyond Computer Science to principally model and improve real-world processes.

If you think you are working on a problem that I might find interesting, do not hesitate to reach out to me to chat about it!

Email
davinchoo [at] seas.harvard.edu
Links
To learn more about me, see here.

Publications

My peer-reviewed publications listed below are tagged and filterable. Note that author names are sometimes in alphabetical ordering of last names as this is the convention for theory work/venues.
Title Authors Venue Year Links
Legend (venues in alphabetical ordering): AAAI: AAAI Conference on Artificial Intelligence; AISTATS: International Conference on Artificial Intelligence and Statistics; ALT: International Conference on Algorithmic Learning Theory; CLeaR: Conference on Causal Learning and Reasoning; COLT: Conference on Learning Theory; DATE: Proceedings of Design, Automation, and Test in Europe; DISC: International Symposium on Distributed Computing; ICML: International Conference on Machine Learning; MLMM: Machine Learning for Molecules and Materials; NeurIPS: Conference on Neural Information Processing Systems; ORL: Operations Research Letters; SODA: Symposium on Discrete Algorithms; UAI: Conference on Uncertainty in Artificial Intelligence

Other writings

Title Remarks Links
Approximate Proportionality in Online Fair Division Joint work with Winston Fu, Derek Khu, Tzeh Yuan Neoh, Tze-Yang Poon, Nicholas Teh arXiv link coming soon
Learning Probabilistic and Causal Models with(out) Imperfect Advice My PhD thesis! PhD Thesis
A short note about the learning-augmented secretary problem Joint work with Chun Kai Ling arXiv
Template for NUS SoC Thesis (and Thesis Proposal) - Overleaf (Read-only)
Uncovering Causal Relationships Using Adaptive Interventions General audience research article AISG ConnectAI article
Template for NUS SoC QE talk slides - Overleaf (Read-only)
Template for NUS SoC QE report - Overleaf (Read-only)
Scribe notes for entire course Massively Parallel Algorithms (Spring 2019)
Lecturer: Mohsen GHAFFARI		 Course webpage, Notes
Scribe notes for entire course Advanced Algorithms (Fall 2018)
Lecturer: Mohsen GHAFFARI		 Course webpage, Notes

Talks

Event Date (DD.MM.YYYY) Topic / Talk title Location / Remarks Links
NUS AlgoTheory Seminar 28.08.2025 Principled AI for Real-world Impact: Structured Decision-Making under Uncertainty NUS, Meeting Room 24 @ COM3 (COM3-02-64) Slides (short)

Slides (long)
Connect and Engage with NUS 2025 26.08.2025 Principled AI for Real-world Impact: Structured Decision-Making under Uncertainty NUS, Multi Purpose Hall (MPH), COM3-01
NTU College of Computing and Data Science (CCDS) Research Seminar 20.08.2025 Principled AI for Real-world Impact: Structured Decision-Making under Uncertainty NTU, LT16, NS1-04-05
SMU School of Computing and Information Systems (SCIS) Research Seminar 18.08.2025 Principled AI for Real-world Impact: Structured Decision-Making under Uncertainty SMU, SCIS 1, Level 5, Meeting Room 5-1
PhD Defence 13.01.2025 Learning Probabilistic and Causal Models with(out) Imperfect Advice Zoom Slides
DSO
technical sharing		 03.10.2024 Algorithms for Learning Probabilistic and Causal Models with Possible Imperfect Advice DSO Playground Slides
Doctoral Seminar 09.09.2024 Algorithms for Learning Probabilistic and Causal Models with Possible Imperfect Advice NUS, SR12, COM3 01-21 Slides
Workshop on Learning-Augmented Algorithms 19.08.2024 Online bipartite matching with imperfect advice (Lightning talk) TTIC, Chicago, IL Paper, Slides
NUS AlgoTheory Seminar 10.06.2024 Online bipartite matching with imperfect advice NUS, MR-20 @ COM3 (COM3-02-59) Paper, Slides
NUS AlgoTheory Seminar 08.04.2024 Envy-free house allocation with minimum subsidy NUS, MR-20 @ COM3 (COM3-02-59) Paper, Slides
Guest presentation at CS6235 03.04.2024 Envy-free house allocation with minimum subsidy NUS, SR@LT19 Paper, Slides
Divesh's research group weekly seminar 15.03.2024 Online bipartite matching with imperfect advice NUS, COM3-02-70. Whiteboard talk -
MPI EI Tea talks 10.08.2023 Recovering causal graphs with adaptive interventions MPI, N 4.022 Slides
NUS AlgoTheory Seminar 17.04.2023 Learning causal DAGs using adaptive interventions NUS, Seminar Room @ LT19 (BIZ 2) Slides
NUS SoC AlgoTheory Group Meeting 24.03.2023 Solving problems using imperfect advice NUS, COM3-02-59 Slides
CS6235 Paper Presentation 08.03.2023 Partitioning Friends Fairly NUS, LT19 Seminar Room Paper, Slides
Computing Research Week - Open House 2023 24.02.2023 Learning Causal DAGs using Adaptive Interventions NUS, Multipurpose Hall 1 (COM3-01-26) Slides
CS6220 Paper Presentation 02.02.2022 Triad Constraints for Learning Causal Structure of Latent Variables Zoom talk and discussion Paper, Slides
CS6101 Paper Presentation 03.09.2021 Online Algorithms with Advice: A Survey Zoom talk and discussion Paper, Slides
Aalto CS Theory Seminar 29.07.2020 k-means++: few more steps yield constant approximation Zoom talk arXiv
MADZ Group Meeting talk 04.05.2020 k-means++: few more steps yield constant approximation Zoom talk arXiv
Reading Group on Discrete and Distributed Algorithms 23.05.2019 Dynamic Algorithms for the Massively Parallel Computation Model ETH. Whiteboard talk. Attached are some pictures. The paper talks about maintaining an approximate MST but I think an exact MST should be doable. See write-up for a sketch. Update: There's a SPAA 2020 paper related to this! arXiv, pic1, pic2, pic3, pic4, pic5
DSO
technical sharing		 02.02.2017 2^{To be, or not to be?}: A look at boolean satisfiability DSO. State-of-the-art methods building upon DPLL and CDCL are covered. An alternative solving method (Stalmarck's method) is also discussed. Animations and some slides removed. Slides
DSO
technical sharing		 10.11.2016 A gentle introduction to community detection DSO. Common methods such as graph partitioning and spectral clustering are discussed. Talk is mainly based off a survey by Santo Fortuno. Animations and some slides removed. Slides

Teaching / Outreach

Place Course Year Links
National University of Singapore (NUS) GET1031 / GEI1000 (Cross listed): Computational Thinking
Lecturers: LEONG Hon Wai and LEOW Wee Kheng 		AY 2021/2022 Sem 1 NUSMODS
National University of Singapore (NUS) CS3230: Design and Analysis of Algorithms
Lecturer: LEE Wee Sun 		AY 2015/2016 Sem 1 NUSMODS
National University of Singapore (NUS) GET1031: Computational Thinking
Lecturers: LEONG Hon Wai and LEOW Wee Kheng 		AY 2015/2016 Sem 2 NUSMODS
National University of Singapore (NUS) CS2020: Data Structures and Algorithms (Accelerated)
Lecturer: Seth GILBERT		 AY 2014/2015 Sem 2,
AY 2015/2016 Sem 2 		NUSMODS
National University of Singapore (NUS) CS1101S: Programming Methodology
Lecturers: Martin HENZ and LOW Kok-Lim 		AY 2014/2015 Sem 1 NUSMODS
National University of Singapore (NUS) CS1231: Discrete Structures
Lecturers: Stéphane BRESSAN and Bryan Kian Hsiang LOW 		AY 2014/2015 Sem 1 NUSMODS
Temasek Junior College (TJC) Initialized and conducted a 12-week student outreach course on Computer Science Jan 2014 - May 2014 Some old, partial
teaching material

Service

  • Reviewer for International Conference on Artificial Intelligence and Statistics (AISTATS), 2026
  • Reviewer for Innovations in Theoretical Computer Science (ITCS), 2026
  • Program Committee for AAAI Conference on Artificial Intelligence (AAAI) Social Impact Track, 2026
  • Program Committee for AAAI Conference on Artificial Intelligence (AAAI) Main Track, 2026
  • Reviewer for Conference on Neural Information Processing Systems (NeurIPS), 2025; Top reviewer
  • Reviewer for Transactions on Machine Learning Research (TMLR), 2025
  • Reviewer for International Conference on Machine Learning (ICML), 2025
  • Reviewer for International Joint Conference on Artificial Intelligence (IJCAI), 2025
  • Reviewer for Conference on Neural Information Processing Systems (NeurIPS), 2024; Top reviewer
  • Reviewer for International Conference on Machine Learning (ICML), 2024
  • Reviewer for International Joint Conference on Artificial Intelligence (IJCAI), 2024
  • Reviewer for International Conference on Artificial Intelligence and Statistics (AISTATS), 2024
  • Subreviewer for Innovations in Theoretical Computer Science (ITCS), 2024
  • Reviewer for Conference on Neural Information Processing Systems (NeurIPS), 2023; Top reviewer
  • Subreviewer for Symposium on Theory of Computing (STOC), 2023
  • Subreviewer for International Colloquium on Automata, Languages, and Programming (ICALP), 2023
  • Reviewer for International Conference on Artificial Intelligence and Statistics (AISTATS), 2023
  • Subreviewer for Conference on Learning Theory (COLT), 2022
  • Subreviewer for Scandinavian Symposium and Workshops on Algorithm Theory (SWAT), 2022
  • Subreviewer for European Symposium on Algorithms (ESA), 2020

For fun

Remarks Links
Recommendations for visiting Singapore I have had given recommendations to various people on multiple occasions so I thought it would be nice to collate it somewhere so that I can just point to a link in the future :) Internal link
IBM Ponder This puzzles Solutions to some IBM's monthly Ponder This puzzles that I attempted. Github
Neopets Shapeshifter solver Used this game as a practice while learning about constraint solvers.
Wrote a JavaScript to extract the puzzle from Neopets, which can be solved by either of my 2 solvers --- One using Google's ortools, one using the MiniZinc constraint solver.		 Directory
Telegram Chess Bot Learnt about the existence of Telegram bots worked. Hacked up a chess bot for fun.
Note: It requires server hosting to work. 		Github
Threshold Secret Sharing Schemes Explored and implemented 3 secret sharing schemes. Github
Cryptopals Solutions to some of the Cryptopals challenges that I attempted. Github
Cipher encodings A collection of cipher encodings. Github

About me

Postdoc
Computer Science
Teamcore
Harvard University
Postdoc advisor: Milind TAMBE

PhD
Computer Science
PhD Thesis

National University of Singapore (NUS)
PhD advisor: Arnab BHATTACHARYYA
PhD co-advisor: Seth GILBERT

Masters
Computer Science
Eidgenössische Technische Hochschule Zürich (ETH Zürich)
Thesis advisor: David STEURER

Undergraduate
Computer Science
First Class Honours
NUS University Scholars Programme (USP)
NUS SoC Turing Programme
National University of Singapore (NUS)
Thesis advisor: Seth GILBERT
UROP advisor: LEE Wee Sun

Undergraduate
Applied Mathematics
First Class Honours
NUS University Scholars Programme (USP)
National University of Singapore (NUS)
Thesis advisor: Frank STEPHAN
Awards (that I am grateful for receiving) Visits / Experiences / Opportunities (which I was fortunate to have)

Short bio (for talks, etc): Davin is a postdoctoral fellow at Teamcore, Harvard University. He earned his PhD in Computer Science from the National University of Singapore (NUS) as an AISG PhD fellow, a Master's degree in Computer Science from ETH Zürich, and two undergraduate degrees in Computer Science and Applied Mathematics from NUS. Between his undergraduate and Masters, he also worked for a while as an applied research scientist at DSO National Laboratories on projects that lie in the intersection of AI and security. During his PhD at NUS, he focused on the foundations of AI and machine learning, working on statistical models, causal inference, and the design of resource-efficient algorithms. His current postdoctoral research at Harvard explores how principled algorithmic and AI techniques can be applied to real-world problems with the goal of achieving meaningful social impact.

Here is my resume (Last updated: November 2025).