Aryaman Jeendgar

Hi! I am Aryaman Jeendgar, a final year undergrad from BITS Pilani, Hyderabad Campus double majoring in Physics and Electronics and Communications engineering.

Broadly, I am interested problems in Optimization Theory and their exciting intersection with modern Deep Learning, and love writing code to serve solutions to such problems!

I will be pursuing the second half of my undergraduate thesis at CERN where my visit will be generously funded by The Department of Physics, Princeton University. I plan on working with Dr. Kilian Lieret on problems concerning a data-driven pipeline for charged particle trajectory prediction and parallely, work with Dr. Henry Schreiner on building scikit-build-core. If you're around in Geneva starting Jan 2024 (or close by in the EU), HMU!

I successfully finished the first half of my undergraduate thesis at the International Computer Science Institute, UC Berkeley where I worked with Dr. Riley J. Murray on problems related to the Operator Relative Entropy Cone and made more technical contributions to the CVXPY codebase in the process! You can find my thesis here

During this time I also simultaneously worked as a Research Intern @ TCS Research with Prof. Mayank Baranwal and Dr. Kushal Chakrabarti on problems related to second-order neural network training dynamics, and analyzing them from an ODE-perspective!

I spent the summer of '23 as a Research Engineer Fellow @ Princeton University (under Princeton Research Computing), where I worked with Dr. Henry Schreiner, contributing to the ongoing scikit-HEP effort. I also participated in the GSoC-2023 program, where I (once again!) contributed to CVXPY!

I spent the summer of '22 writing open source code for CVXPY, under the GSoC-2022 program and interning at Intel Labs in the Cloud Systems Research group, where I worked with Dr. Sameh Gobriel on online optimization approaches to scaling out Nearest Neigbour Search queries for graph databases.

In the past, I worked with Professor Snehanshu Saha (BITS Goa) and Mr. Soma S. Dhavala (Wadhwani AI) on problems related to quantile regression in the context of neural networks and systematic methods for accelerating training for such scenarios, a pre-print of our latest work can be found HERE