RESEARCH / TEACHING / ENTREPRENEURSHIP / INTERNSHIPS
Research & experience.
Research in multimodal medical imaging, a foundation in language and graph learning, and experience translating technical ideas through teaching and assistive technology.
CURRENT RESEARCH FOCUS
Multimodal segmentation for quantitative theranostic imaging.
My current work combines CT anatomy, SPECT uptake, and clinical report information to localize and segment organs and tumors in post-therapy Lu-177 DOTATATE/PSMA imaging.
Alongside the methods research, I am developing TheraSeg to bring segmentation, visualization, quantitative analysis, and structured export into a research software platform.
Multimodal medical imaging, unsupervised learning, language, and graph-based models.
CURRENT ROLE
Graduate Research Assistant
Washington University in St. Louis
Computational Medical Imaging and Therapy (CMIT) Lab
St. Louis, Missouri, USA
– Present
Advisor: Abhinav K. Jha, Ph.D.
Developing multimodal organ and tumor segmentation methods for post-therapy Lu-177 SPECT/CT, with applications to quantitative imaging, dosimetry, and treatment-response assessment.
Medical imaging
Multimodal AI
Unsupervised learning
Clinical text processing
Research software
Contributions & work
Integrating CT-derived anatomy, SPECT uptake patterns, and LLM-extracted information from clinical radiology reports to guide tumor localization and segmentation.
Developing unsupervised segmentation workflows that combine multimodal imaging information with clinical knowledge.
Developing TheraSeg, a research software platform combining segmentation, multiplanar visualization, quantitative analysis, and structured data export.
Presented research and software at the 2026 SNMMI Annual Meeting and SNMMI AI Summit.
Computational Medical Imaging and Therapy (CMIT) Lab
St. Louis, Missouri, USA
–
Advisor: Abhinav K. Jha, Ph.D.
Investigated unsupervised segmentation of noisy SPECT images using hidden Markov random fields and expectation maximization.
SPECT imaging
HMRF-EM
Unsupervised segmentation
Denoising
MATLAB
Contributions & work
Developed an HMRF-EM method for segmentation of noisy SPECT images.
Assessed segmentation with Dice similarity and investigated the effect of Noise2Void denoising.
Deep Learning Researcher
LBR Research Labs
India · Remote
–
Studied inductive graph neural networks for document classification and graph-based approaches to hate-speech classification.
Achieved a micro-F1 of 0.61 on hate-speech classification.
Graph neural networks
Inductive learning
Graph attention
NLP
Text classification
Model evaluation
Contributions & work
Conducted the inductive GNN text-classification study during this role, comparing TexTING, inductive GCN, and inductive GAT on AG News, IMDB, and DBpedia.
Co-authored “Comparative Study of Inductive Graph Neural Network Models for Text Classification,” published at IEEE ASSIC 2022.
Also evaluated multi-level graph neural networks for hate-speech classification.
Implemented a graph-attention formulation achieving a micro-F1 score of 0.61.
Developed an unsupervised word-sense-disambiguation pipeline for low-resource statistical machine translation.
Word sense disambiguation
Low-resource NLP
Machine translation
Unsupervised learning
Contributions & work
Integrated unsupervised word sense disambiguation into a statistical machine translation workflow, improving English-to-Hindi translation by 2.3 BLEU points.
Co-authored the related research published in IETE Journal of Research.
Imaging mathematics, linear algebra, and practical image processing.
Teaching Assistant
Washington University in St. Louis
Department of Biomedical Engineering
St. Louis, Missouri, USA
–
SUBJECT
Mathematics of Image Science
BME 470/570 · ESE 5931
Teaching assistant for Mathematics of Image Science in the Department of Biomedical Engineering.
Imaging mathematics
Teaching
Teaching Assistant
Washington University in St. Louis
Summer Mathematics Crash Course
St. Louis, Missouri, USA
–
SUBJECT
Linear Algebra
Teaching assistant for the Linear Algebra component of WashU’s Summer Mathematics Crash Course.
Linear algebra
Mathematics teaching
Teaching Assistant
National Institute of Technology Hamirpur
Hamirpur, Himachal Pradesh, India
–
SUBJECT
Digital Image Processing Laboratory
Teaching assistant for the Digital Image Processing Laboratory.
Digital image processing
Laboratory teaching
03 / EXPERIENCE
Entrepreneurship
Assistive technology and the translation of computer vision into products.
Co-Founder & Chief Technology Officer
Oculosense EdgeAi Devices
India
–
Co-founded Oculosense and led technical development of AI-powered assistive devices combining computer vision and hardware.
Computer vision
Hardware integration
Assistive devices
Technical leadership
Product development
Contributions & work
Co-founded the company to develop assistive technology for people with visual impairments.
Led technical development of AI-powered assistive devices, combining computer vision with hardware.
Participated in the NSRCEL–Texas Instruments Launchpad program; the team was selected among the top 20 nationwide and secured product-development support, mentorship, and DST seed funding.