Poojan Shah

About

I am a 3rd year undergraduate student at the Department of Computer Science & Engineering at IIT Delhi . I will be spending the winter semester of 2025 as an exchange student at the Faculty of Mathematics at the University of Waterloo.

I am excited to join Atlas Research as a quantitative trading and research intern during the summer of 2025. I spent the summer of 2024 at the Computer Science Group of the Center for Quantum Technologies at the National University of Singapore as a research intern.

[Me] and my brief [CV] (last updated on Feb 6 2025). [Google Scholar].

Research

I enjoy thinking about problems broadly under theoretical computer science including algorithms, cryptography, quantum information processing and statistical learning.

I am currently studying methods inspired from statistical physics for a class of problems in statistical learning referred to as planted inference, for example the Continuous Learning With Errors problem from Lattice based cryptography.

During much of 2024, I studied an algorithmic framework for efficient classical simulation of quantum machine learning algorithms in the QRAM model referred to as Dequantized QML and worked on applying it to sampling based algorithms for clustering problems.

During my time at CQT, I studied pseudo random quantum states which are a potentially weaker cryptographic assumption compared to one-way functions.

Publications

  • Poojan Shah and Ragesh Jaiswal (2025). Quantum (inspired) \(D^2\)-sampling with applications. In The Thirteenth International Conference on Learning Representations. [OpenReview][arXiv]

Preprints

  • Poojan Shah, Shashwat Agrawal and Ragesh Jaiswal (2025). A new rejection sampling approach to k-means++ with improved trade-offs. [arXiv] [code]

Miscellaneous

Contact

IITD e-mail : \( \mathtt{cs1221594@iitd.ac.in}\)

Personal e-mail : \( \mathtt{shahpoojan2004@gmail.com} \)