Pawan Kumar

About Me

I have done my PhD and postdocs from Europe that gave me strong foundations in research. For the past 5 years, I have been working as a full-time data science researcher at IIIT-Hyderabad, India. I have more than 10 years of hands-on research experience in Machine learning, Optimization, High performance computing, and applications to sciences.

I have guided 8 thesis students and am currently guiding 10. I have obtained numerous industrial grants from Microsoft, Qualcomm, and TATA. I like working on all aspects of machine learning: theoretical, algorithmic and applied. I consistently publish in top venues like AAMAS, SIAM J. of Optimization, Numerische Mathematik, SDM, ECML, WACV, and NeurIPS-W.

Employment

• ResearcherNow
  IIIT, Hyderabad, India
• Postdoc, FU Berlin, Germany
• Postdoc, Fraunhofer ITWM, Kaiserslautern, Germany
• Postdoc, KU Leuven and exascience lab, Belgium

Education

• PhD (Highest honours ("très honorable" in French))
  INRIA, France
• MS
  Indian Institute of Technology, Guwahati, India

Computer Skills

Publications

Books

• Mathematics of Generative Modelling, P. Kumar, et al., In preparation, 450 pages, 2024.

Safety and Alignment of AI

• MA-SafeDiffuser: Safe Multi-Agent Planning with Diffusion Probabilistic Models, K. Ravish, et al., AAMAS 2026.
• Safe Offline Reinforcement Learning using Diffusion Policies, A. Kushwaha, et al., AAMAS 2026.
• Alignment and Safety of Diffusion Models via Reinforcement Learning and Reward Modeling: A Survey, P. Lamba, et. al., under submission, arXiv 2025.
• A Survey of Safe Reinforcement Learning and Constrained MDPs, A. Kushwaha, et. al., under submission, arXiv 2025.

Large Language Models

• Hierarchical Sparse Plus Low Rank Compression of LLM, Aditi Gupta, et. al., CODS 2025.

Multimodal AI

• Fast and Efficient Modern BERT based Text-Conditioned Diffusion Model for Medical Image Segmentation, Siddharth, et. al., CVIP 2025.

Machine Learning for Sciences: Biosciences, Circuit Design, etc

• A label-free sensing of creatinine using radio frequency-driven lab-on-chip (loc) system, S. Sinha, et al., Engineering Express Journal 2024.
• A Label-Free Low-Cost Radio Frequency Driven Noninvasive Lab-on-Chip System for Creatinine Detection, S. Sinha, et al., ICST 2024.
• Qualitative Data Augmentation for Performance Prediction in VLSI circuits, P. Srivastava, P. Kumar, Z. Abbas, Elsevier, VLSI, Integration Journal, 2024.
• Data Augmentation for Performance Prediction in VLSI circuits, P. Srivastava, P. Kumar, Z. Abbas, ISCAS, Singapore 2024.
• Enhancing ML model accuracy for Digital VLSI circuits using diffusion models, P. Srivastava, Z. Abbas, P. Kumar, MLSys, NeurIPS, 2021.

Data Mining and Machine Learning

• alphaElimination: Using deep reinforcement learning for sparse Gaussian Elimination, A. Dasgupta, P. Kumar, ECML 2023.
• marl-jax: Multi-agent Reinforcement Leaning framework for Social Generalization, K. Mehta, A. Mahajan, P. Kumar, ECML 2023.
• Effects of Spectral Normalization in Multi-agent Reinforcement Learning, K. Mehta, A. Mahajan, P. Kumar, IJCNN 2023.
• (Best Paper Award) Light Weight Deep Extreme Multilabel Classification, U. Mishra, et al., IJCNN 2023.
• A Riemannian Approach to Extreme Classification Problems, J. Naram, T. Sinha, P. Kumar, CODS-COMAD, 2022.
• Hybrid Tokenization and Datasets for Solving Mathematics and Science Problems Using Transformers, P. Mandlecha, et al., SDM, 2022.
• SCIMAT: Science and Mathematics Dataset, S. Chatakonda, et al., DCAI, NeurIPS-W, 2021.
• DXML: Distributed Extreme Multilabel Classification, Springer, P. Kumar, BDA 2021.

Computer Vision

• Fast and Efficient Modern BERT based Text-Conditioned Diffusion Model for Medical Image Segmentation, Siddharth, et. al., CVIP 2025.
• Angle based dynamic learning rate for gradient descent, N. Mishra, P. Kumar, IJCNN 2023.
• Adaptive Concencous Optimization for GANs, S. Danisetty, et al., IJCNN 2023.
• Nonnegative Low-Rank Tensor Completion via Dual Formulation with Applications to Image and Video Completion, T. Sinha, et al., WACV, 2022.
• Structured Low-Rank Tensor Learning, J. Naram, T. Sinha, P. K., NeurIPS-W, 2021.
• A Deflation Based Fast and Robust Preconditioner for Bundle Adjustment, S. Das, et al., WACV 2021.
• Multigrid Preconditioned Solver for Bundle Adjustment, S. Katyan, et al., WACV 2020.

Optimization Methods and Preconditioners

• Riemannian Hamiltonian methods for min-max optimization on manifolds, A. Han, et al., SIAM J. of Optimization, SIOPT, 2023.
• Fast Preconditioned Solver for Truncated Saddle Point Problem in Nonsmooth Cahn-Hilliard Model, Book chapter, 2016.
• Modified tangential frequency filtering decomposition and its Fourier analysis, Numerische Mathematik, 2010.

High Performance and Scientific Computing

• Efficient FPGA Implementation of Conjugate Gradient Methods for Laplacian System using HLS, FPGA 2019.
• Multilevel Communication Optimal Least Squares Solver, ICCS, vol. 51, 2015.
• High Performance Solvers for Implicit Particle in Cell Simulation, ICCS, vol. 18, 2013.

Grants & Awards

• • KCIS Grant (TATA group): ML for biosensors
• • Microsoft Academic Partnership Grant
• • Best Paper Award, IJCNN 2023
• • Qualcomm Faculty Award
• • Marie Curie Actions Fellowship
• • GATE Percentile 96.6 (Rank 88), 2007

Teaching History

Selected Talks & Workshops

• A Fast and Efficient Modern BERT based Text-Conditioned Diffusion Model for Medical Image Segmentation, CVIP, IIT Ropar, December 2025
• Hierarchical Sparse Plus Low Rank Compression of LLM, CODS, IISER, Pune, December 2025
• Effects of Spectral Normalization in MARL, Gold Coast, Australia, 2023
• Parallel aggregation based algebraic multigrid, Domain Decomposition, France, 2012
• Winter school on Hierarchical matrices, Leipzig, Germany, 2014
• Research visit (Prof. M. Gander) University of Geneva, 2014