Agam Chopra

Utah Teapot!

I am a results-driven AI researcher with a strong foundation in computer vision, artificial intelligence, and deep learning, complemented by hands-on experience in generative AI and multi-modal data integration. My work focuses on solving real-world problems through innovative AI solutions, including cross-modal 3D medical image synthesis, AI-driven MRI optimizations, and efficient algorithm development for clinical applications.
I bring extensive experience in building scalable, high-performance AI systems, including developing open-source tools, designing data pipelines, and creating production-ready solutions. My expertise spans Python, PyTorch, CUDA, and high-performance computing, with a proven track record of collaborating across disciplines to deliver impactful projects.
With multiple publications under my belt, I excel in translating cutting-edge research into practical, industry-relevant applications. I’m passionate about leveraging AI to drive innovation and efficiency, whether through developing next-generation models, optimizing workflows, or mentoring teams to achieve technical excellence.
I love Cats, Physics, and Computer Science. You can find me tinkering around with robotics and fun programming side projects in my free time.

For any business inquiries or hiring queries, please contact me via my work email linked above. To request a full up-to-date resume, please add "resume-request" to the email heading.

I am based in Chicago, IL and currently open to full-time, part-time, contract, and consultant opportunities (remote, hybrid, or in-person). Base annual salary expectation of $150,000, negotiable.

Work Experience

Doctoral Student
KurtLab, University of Washington, Seattle, WA, USA
Sep 2022 - present

• Created open-source packages and tools for easy integration of medical data processing pipelines with gradient friendly PyTorch implementations for the medical AI community
• Experience working with multi-modal datasets of image, 3D scans, temporal scans, text, and other tabular metadata for training cross-modal generative AI algorithms
• Participated and achieved competitive scores in 2022, 2023, and 2024 Brain Tumor Sequence challenges, for 3D image synthesis, tumor classification, and pre-post operative registration
• Researching novel Generative AI frameworks for cross-modal 3D medical image synthesis, focusing on Synthetic Tau-PET and Compression-Synthesis techniques to streamline image synthesis for clinical applications
• Successfully obtained funding from competitive Amazon research grant and internal Data Science grant at the University of Washington for investigating novel data-driven research proposals
• Collaborated on Amplified 4D MRI and AFLOW projects to enhance iterative physics algorithms with the aid of artificial intelligence, optimizing computational efficiency

Graduate Teaching Assistant
University of Washington, Seattle, WA, USA
Jan 2023 - present

• Instructed 1 quarter of ME535 and 3 quarters of ME230 to a class of 200 students, focusing on application of computational techniques in engineering by delivering lectures and hands-on lab sessions in Python, C, and C++, emphasizing algorithm development and problem-solving techniques
• Taught core mathematical concepts, including numerical methods, linear algebra, and differential equations, foundational to the courses

Research Assistant
KurtLab, Stevens Institute of Technology, Hoboken, NJ, USA
Jan 2021 - Jan 2022

• Researched, tested, and implemented deep learning-based 4D spatial-temporal image registration algorithms to accurately extrapolate displacement fields of 3D aMRI time sequences using Python and MATLAB
• Theorized a novel deep learning-based automatic brain morphology algorithm to accurately estimate the morphology of the target region of a brain scan for any patient sample

Software Developer I, II
Reindeer Shuttle, INC, West Lafayette, IN, USA
May 2018 - Jun 2020

• Optimized the company's website using Google Analytics, PWA, and UI design principles
• Conducted data analysis to identify trends and correlations for cost-effective fleet operations
• Presented data-driven reports using Google Analytics, Python, and Excel

Education

Ph.D. in Mechanical Engineering (Deep Learning & Computer Vision)
University of Washington, Seattle, WA (2022 - present)

M.S. in Computer Science (Machine Learning)
Stevens Institute of Technology, Hoboken, NJ (2020 - 2022)

B.S. in Physics
Purdue University, West Lafayette, IN (2015 - 2020)

Publications

1. Chopra, A., Ren, T., Rivera, J. E. H., Jahanian, H., Hedden, T., & Kurt, M. (2024). Synthesizing Tau-PET from multimodal MRI for Alzheimer’s disease detection using disentangled quantized variational autoencoders. In Medical Imaging with Deep Learning.
2. Chopra, A. S., Heras Rivera, J. E., Ren, T., Oswal, H., Pan, Y., Sordo, Z., Walters, S., Henry, W., Mohammadi, H., Olson, R., Rezayaraghi, F., Lam, T., Jaikanth, A., Kancharla, P., Ruzevick, J., Ushizima, D., & Kurt, M. Medical Image Inpainting using Efficient Transformer and Fourier UNet. Paper submitted for BRATS 2024.
3. Ren, T., Rivera, J. E. H., Chopra, A., Ruzevick, J., & Kurt, M. (2024). Here comes the explanation: A Shapley perspective on multi-contrast image segmentation. In Medical Imaging with Deep Learning.
4. Rivera, J. E. H., Kurt, M., Ren, T., Chopra, A., Oswal, H., & Pan, Y. (2024). How we won the ISLES 2024 Challenge by windowing and preprocessing. In Medical Imaging with Deep Learning.
5. Ren, T., Honey, E., Rebala, H., Sharma, A., Chopra, A., & Kurt, M. (2023). An optimization framework for processing and transfer learning for the brain tumor segmentation. In International Challenge on Cross-Modality Domain Adaptation for Medical Image Segmentation (pp. 165–176). Springer Nature Switzerland Cham.
6. Ren, T., Sharma, A., Rivera, J. E. H., Rebala, L. H., Honey, E., Chopra, A., & Kurt, M. (2024). Re-DiffiNet: Modeling discrepancy in tumor segmentation using diffusion models. In Medical Imaging with Deep Learning.
7. Javid Abderezaei, Pionteck, A., Chopra, A., & Kurt, M. (2024). 3D Inception-Based TransMorph: Pre- and post-operative multi-contrast MRI registration in brain tumors. In S. Bakas et al. (Eds.), Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. BrainLes 2022. Lecture Notes in Computer Science (Vol. 14092, pp. 35–45). Springer, Cham.
8. Rivera, J. E. H., Chopra, A. S., Ren, T., Oswal, H., Pan, Y., Sordo, Z., Walters, S., Henry, W., Mohammadi, H., Olson, R., et al. (2024). An ensemble approach for brain tumor segmentation and synthesis. arXiv preprint arXiv:2411.17617.

Relevant Projects

• Reinforcement Learning-Based Autonomous Video Game Agent (ongoing)
  Designed innovative Reinforcement Learning-driven Artificial Intelligence to autonomously master First-Person video games


• Bio-Inspired Artificial Neural Networks for high dimensional Multi-modal Data Stream (ongoing)
  Working on backpropogation free high sample efficiency framework for Artificial Neural Networks that uses optimization and growth algorithms inspired form biological mechanisms in early brain development.


• Synthetic Image Synthesis using Diffusion-based A.I. (complete)
  Coded Stable diffusion based generative efficient attention models to synthesize realistic images from noise, text, and images


• Synthetic Image Synthesis using WGAN-GP-based A.I. (complete)
  Programmed generative model based on Wasserstein GAN with gradient penalty to synthesize images from pure noise


• 3D Real-time N-body Newtonian Particle Dynamics Physics Simulator (complete)
  Contrived real time 3D sandbox simulation toolbox for simulating N-body dynamics on multicore CPUs and CUDA GPUs


• Torch-Register, Torch-NMI, 3D Edge Loss, and Standardized 3D Medical Image Processing Pipeline for PyTorch (complete)
  Developed and published open-source 3D and 2D medical image processing software for Python3 and PyTorch on Pypi.org


• 3D-Scene Generation from 2D Images using Conventional and A.I. Algorithms (complete, continuous update)
  Wrote applications to transfer 2D image domain to approximate 3D scene domain using AI and conventional mathematics


• A.I. based Pet Companion Robot (complete, revision ongoing)
  Prototyped robots to recognizing pets and avoiding collisions with obstacles using deep learning-based object detection


• Life Engine based on my 3D Physics Engine and Evolutionary Reinforcement Learning (ongoing)
  Developed real-time AI simulations for complex organism evolution in dynamic environments. Implemented intricate physical, sensory(visual, tactile, thermal, and chemical interaction models, enabling organisms to adapt and compete effectively