About Me

I am a researcher specializing in accelerated Magnetic Resonance Imaging (MRI) acquisition and reconstruction, compressed sensing, and deep learning for biomedical applications. I work in the Inria-CEA joint MIND team at NeuroSpin (CEA Paris-Saclay) and Inria, designing non-Cartesian trajectories and learning-based reconstruction pipelines to push the spatio-temporal boundaries of high-field MRI scans.

Background & Education

I trained as an electronics and communication engineer at the National Institute of Technology Karnataka, Surathkal, completing a B.Tech in 2018 (CGPA 9.96/10). I then spent a year at NVIDIA in Bangalore as an architect in the GPU Memory Subsystem group, building and testing functionally accurate and performance models of NVLink.

In 2019 I moved to France and joined NeuroSpin as a research engineer, applying compressed sensing to the acquisition and reconstruction of Magnetic Resonance Imaging and Scanning Electron Microscopy data. I was granted equivalence to a Master’s degree in December 2019 and enrolled in the Electrical, Optical, Bio-Physics and Engineering doctoral school at Université Paris-Saclay.

From November 2020 to April 2023 I was a NUMERICS fellow and doctoral student at NeuroSpin, working with the Inria MIND team. My thesis, Designing and learning non-Cartesian k-space sampling trajectories for accelerated 3D MRI, treats the sampling pattern itself as a variable to be optimized: the trajectory design and the reconstruction network are learned jointly rather than tuned one after the other, so that high-resolution 3D scans at 7T and 11.7T can be acquired substantially faster, and with fewer artifacts. Since November 2023 I have been a researcher with the same team, working on accelerated acquisitions and artifact-free reconstructions.

Current Work

My current research focuses on developing open-source computational tools and advanced reconstruction algorithms for neuroimaging and body MRI. Key ongoing efforts involve:

  • Designing hardware-compliant, optimized k-space trajectories to minimize scan times while maximizing signal-to-noise ratio.
  • Physics-guided neural networks and unrolled reconstruction architectures capable of handling highly undersampled multi-coil MRI data.
  • Reproducible imaging pipelines and open-source packages for the broader MR research community.

Research Interests

My primary research interests span:

  • Accelerated MRI Acquisition & Reconstruction
  • Non-Cartesian Trajectory Design (SPARKLING, 3D non-Cartesian imaging)
  • Deep Learning for Inverse Problems in Imaging
  • Compressed Sensing & Optimization
  • Ultra-High Field Neuroimaging
  • Scientific Computing & Open Science