Shamak Dutta
I am a research scientist at Meta where I work on large scale problems in machine learning and optimization.
I was a postdoctoral researcher in Electrical and Computer Engineering (ECE) at the University of Waterloo where I worked with Yash Pant and Stephen L. Smith on combinatorial optimization problems in flight planning (with Airbus), data-efficient learning in control (PhD thesis), multi-objective optimization and motion planning.
I develop algorithms: exact formulations (linear programming, convex optimization, mixed integer programming), approximation algorithms (worst-case guarantees), and heuristics (fast, high-quality solutions in practice).
I'm also an experienced coder having completed eight internships (software engineering + research) at companies including Amazon in Palo Alto, Latent Logic in Oxford (now Waymo UK), and Preferred Networks in Tokyo.
Teaching
I was a teaching assistant for several courses including Algorithms (ECE 250, ECE 406) and Probability Theory & Statistics (ECE 203, ECE 307). I enjoy teaching and have consistently received high student evaluations through course reviews as well as a teaching assistant award. Here are some examples of student feedback:
- "Just phenomenal, he is extremely clear during tutorials. He has a direct and easy to understand lesson plan that starts with review and finishes with questions. Everything is clearly explained, I think I understood more from the tutorials than lectures." (ECE 307)
- "Hands down the smartest and most knowledgeable and helpful TA I have ever interacted with. Thanks for all the help Shamak and all the best, you are going places!" (ECE 250)
- "Very clear and helpful. Seemed to have a good knowledge of the course material and was helpful when I asked him questions. He is a very clear communicator." (ECE 250)
- "Banger tutorials." (ECE 203)
Publications & Preprints
Efficient Multi-Objective Planning with Weighted Maximization using Large Neighbourhood Search.
Krishna Kalavadia, Shamak Dutta, Yash Pant,
and Stephen L. Smith
IEEE International Conference on Robotics and Automation (ICRA), Austria, 2026.
Hierarchical Informative Path Planning via Graph Guidance and Trajectory Optimization.
Avraiem Iskandar, Shamak Dutta, Kevin Murrant, Yash Pant,
and Stephen L. Smith
IEEE American Control Conference (ACC), New Orleans, USA, 2026.
Informative Path Planning for Active Regression with Gaussian
Processes via Sparse Optimization.
Shamak Dutta, Nils Wilde,
and Stephen L. Smith
IEEE Transactions on Robotics (T-RO), 2025.
A Unified Approach to Optimally Solving Sensor Scheduling and Sensor Selection Problems in Kalman Filtering.
[arXiv]
Shamak Dutta, Nils Wilde,
and Stephen L. Smith
IEEE Conference on Decision and Control (CDC), Singapore, 2023.
Approximation Algorithms for Tours in Random Fields with Guaranteed Estimation Accuracy.
[arXiv]
Shamak Dutta, Nils Wilde, Pratap
Tokekar, and Stephen L. Smith
IEEE International Conference on Robotics and Automation (ICRA), London, UK, 2023.
Informative Path Planning in Random Fields via Mixed Integer Programming. [arXiv]
Shamak Dutta, Nils Wilde, and Stephen L. Smith
IEEE Conference on Decision and Control (CDC), Cancun, Mexico, 2022.
An Improved Greedy Algorithm for Subset Selection in Linear Estimation. [PDF]
Shamak Dutta, Nils Wilde, and Stephen L. Smith
IEEE European Control Conference (ECC), London, United Kingdom, 2022.
Convolutional Neural Networks Regularized by Correlated
Noise. [PDF]
Shamak Dutta, Bryan Tripp, and Graham Taylor.
Conference on Computer and Robot Vision, Toronto (CRV), Canada, 2018.
Barcodes for Medical Image Retrieval Using Autoencoded Radon
Transform. [PDF]
H.R. Tizhoosh, Christopher Mitcheltree, Shujin Zu, Shamak Dutta.
International Conference on Pattern Recognition (ICPR), Cancun, Mexico, 2016.
Theses
Resource Constrained Linear Estimation in Sensor Scheduling and Informative Path Planning.
[PDF]
PhD Thesis, University of Waterloo. 2024.
Correlated Noise in Deep Convolutional Neural Networks.
[PDF]
MASc Thesis, University of Waterloo. 2019.
|
|