CV

Education

Oct 2025 – present
Technische Universität München
Ph.D. in Computer Science
  • Advisor: Prof. Hartwig Anzt
  • Research on randomized numerical linear algebra (RandNLA) and efficient GPU implementations.
Aug 2019 – May 2024
BITS Pilani, India
M.Sc. Physics & B.E. (Hons.) Electrical & Communications Engineering

Research Experience

Aug 2024 – Aug 2025
Machine Learning Research Engineer, International Computer Science Institute, UC Berkeley
Berkeley, California
  • Host: Prof. Michael Mahoney
  • Researching preconditioners for solving large-scale linear systems leveraging randomized sketching.
  • Building the RandBLAS and RandLAPACK libraries as part of the BALLISTIC project.
  • Working on a highly-parallelized, differentiable path integration library, padaquad, and exploring problems in neural network mode-connectivity.
Jan 2024 – Jun 2024
Research Fellow via Princeton University, CERN
Geneva, Switzerland
  • Host: Dr. Peter Elmer; Mentors: Dr. Kilian Lieret, Dr. Gage DeZoort, Dr. Henry Schreiner
  • Worked on extending the GNN-tracking pipeline for charged-particle-tracking in various ways.
  • Explored the introduction of a noise-classifier module and the use of uncertainty quantification techniques.
Sept 2023 – Dec 2023
Research Intern, TCS Research
Mumbai, Maharashtra
  • Host: Prof. Mayank Baranwal
  • Worked on developing new second-order optimizers for deep learning using tools from control and dynamical systems theory.
  • Designed and implemented experiments and assisted in theoretical analyses.
Aug 2023 – Dec 2023
Master's Thesis, International Computer Science Institute, UC Berkeley
Remote
  • Host: Dr. Riley J. Murray
  • Worked on a variety of problems centered around the Operator Relative Entropy Cone and its semidefinite approximation.
  • Introduced dual variables for the implementation of the N-dimensional power cone PowConeND.
  • Implemented an atom for the Quantum Relative Entropy (QREP) and various other quantum information modelling functionality.
June 2023 – Sep 2023
Summer Research Software Engineer Fellow, Princeton Research Computing
Hybrid, Princeton, NJ
  • Mentor: Dr. Henry Schreiner
  • Worked on tools part of the ongoing scikit-HEP project.
  • Implemented serialization spec within HDF5 for the UHI-interface and built a textual powered TUI for copier and cookiecutter projects.
June 2022 – Sep 2022
Graduate Technical Intern, Intel Labs
Cloud Systems Research Lab · Bangalore, Karnataka
  • Manager: Nilesh Jain, collaboration with Dr. Sameh Gobriel
  • Wrote a shard mode of operation for VDMS that linearly scaled out Add queries.
  • Researched online clustering algorithms for optimization of Approximate Nearest Neighbor queries.
Aug 2021 – Feb 2022
Undergraduate Researcher, BITS Pilani
Goa Campus, Dept. of CS
  • With Prof. Snehanshu Saha and Dr. Soma S. Dhavala
  • Developed a novel Quantile Regression based framework around the proposed loss function in the Deep Learning setting.

Publications

  • Randomized Sketching is Robust to Low-Precision Rounding on GPUs Aryaman Jeendgar, Clément Flint, Hartwig Anzt arXiv:2606.20195, 2026 arXiv
  • CVXPY 1.9: Recent Advances in Optimization Modeling Software William Z. Zhang, Parth Nobel, Aryaman Jeendgar, Riley Murray, Philipp Schiele, Steven Diamond arXiv:2606.14891, 2026 arXiv
  • Preconditioning via Randomized Range Deflation (RandRAND) Oleg Balabanov, Caleb Ju, Kaiwen He, Aryaman Jeendgar, Michael W. Mahoney arXiv:2509.19747, 2025 arXiv
  • LogGENE: A smooth alternative to check loss for Deep Healthcare Inference Tasks Aryaman Jeendgar, Aditya Pola, Soma S. Dhavala, Snehanshu Saha arXiv:2206.09333v3, 2022 arXiv

Workshops & Talks

2026-08
Fused GPU Kernels for Two-Sided Randomized Sketching
GAMM Workshop on Applied & Numerical Linear Algebra, 2026 · Magdeburg · Contribution Link
  • Authors: Aryaman Jeendgar, Clément Flint, Kanaya Ozora, Hartwig Anzt
2026-06
Petrov-Galerkin Approximations in (Multi)-Linear Algebra from the Lens of Optimization
2026 SIAM Conference on Optimization (OP26) · Edinburgh · Session MS97
  • Authors: Nicolas Venkovic, Aryaman Jeendgar, Hartwig Anzt
2026-03
Parallel Locally Optimal Iterative Methods
2026 SIAM Conference on Parallel Processing for Scientific Computing · Berlin · Session CP10
  • Authors: Nicolas Venkovic, Amir Bouslama, Aryaman Jeendgar, Hartwig Anzt

Teaching

Summer 2026
CITHN4021 Convex Optimization
Master's level · TUM
  • Designed (from scratch) and was the lead instructor for a master's-level convex optimization course at TUM. Covering convex analysis, first and second order methods, deep-learning optimizers and advanced conic modelling via CVXPY
  • Course material (slides, exercises, projects, final exam --- all with solutions) will be uploaded here soon.

Open Source

2022 / 2023
Google Summer of Code — CVXPY
  • 2022: Implemented the Von Neumann entropy atom and an approximate scalar relative entropy cone via Gauss–Legendre quadrature-based semidefinite approximations.
  • 2023: Added stationarity-of-Lagrangian verification primitives for differentiable CVXPY problems (KKT-style debugging helpers).
2022 – 2023
scikit-HEP & scikit-build-core
  • General build-system and packaging improvements across the scikit-HEP ecosystem, including UHI-interface serialization in HDF5.
  • Built a TUI: CopyCuTTer for the copier and cookiecutter projects

Skills

  • Tools and Languages: Python, NumPy, CVXPY, scikit-learn, PyTorch, C++, CUDA, Git, Emacs, LaTeX
  • HPC / GPU: CUDA, cuBLAS, cuSPARSE, mixed-precision arithmetic
  • Numerical: Randomized NLA (CountSketch, OSNAP, SparseStack), iterative solvers, SDP / SOCP / LP solvers
  • Optimization modeling: CVXPY, disciplined convex programming
  • Tools: Git, CMake, scikit-build-core, Docker, GitHub Actions