Profile

Hello there!

Computer Science graduate from IIT Jammu with R&D research experience in computer vision, medical imaging, and NLP. Worked with Boston College, Yale, UPenn, UoL, and The Captury on projects like cervical cancer dose prediction, brain tumor segmentation, and motion capture analysis. Published at SPIE, CVPR Workshops, CIBM, and MICCAI. Currently working in The Captury GmbH as an AI Research Engineer I.

Publications

Dendrite cross attention for high-dose-rate brachytherapy distribution planning
Saini Sourav, Liu Xiaofeng
Computers in Biology and Medicines, 2025
Paper Link
High-dose-rate Brachytherapy Planning with Dendrite Cross-Attention UNet
Saini Sourav, Li Zhen, Liu Xiaofeng
SPIE Medical Imaging Conference, 2024
Paper Link
NTIRE 2024 Challenge on HR Depth from Images of Specular and Transparent Surfaces
Pierluigi Zama Ramirez, Fabio Tosi, Saini Sourav (contributor)
Proceedings of the IEEE/CVF CVPR Workshop, 2024
Paper Link
Multiscale Encoder and Omni-Dimensional Dynamic Convolution Enrichment in nnU-Net for Brain Tumor Segmentation
Mistry Sahaj, Saini Sourav, Aashray Gupta, Aayush Gupta
MICCAI BrainLesion Workshop, 2023
Paper Link

Experiences

AI Research Engineer I - The Captury

Feb 2025 - Present

I am currently working as an AI research engineer at The Captury, where I am working on facial expression tracking and motion capturing. Working directly with the CTO and Co-Founder on some really cool stuff for Markless motion capture.

PyTorch Computer Vision Deep Learning Facial Expression Tracking

Research Associate - Boston College

April 2025 - June 2025

Worked with Dr. Donglai Wei on finetuning ST-Unet for 3D instance segmentation for microscopy images.

Project Link

PyTorch Medical Imaging Deep Learning UNet Computer Vision Segmentation

Research Associate - University of Liverpool

Jan 2024 - Dec 2024

Worked as a research associate researching the early detection of anorexia through Reddit posts. Using a Time Series Transformer and LLM, I have achieved an accuracy of 85.2%. Also submitted a research paper to the ACL 2025 conference.

Project Link

PyTorch NLP Deep Learning Time Series Analysis LLM Classification

Research Affiliate - Yale University

May 2024 - September 2024

I was the sole undergraduate researcher in the lab, where I developed a novel DCA-UNet architecture aimed at predicting dose maps for internal beam radiation therapy in cervical cancer patients. This work culminated in a publication at the SPIE Medical Imaging Conference 2025, in the AI/ML track.

Project Link

PyTorch Medical Imaging Computer Vision PyTorch SPIE Medical Imaging 2025 Deep Learning

Research Student - UPenn

May 2023 - August 2023

Developed novel brain tumor segmentation model by modifying the existing nn-UNet architecture. I co-authored a research paper at MICCAI BrainLesion Workshop 2023

Project Link

nnU-Net Medical AI Python PyTorch Computer Vision Brain Tumor Segmentation MICCAI

Education

B.Tech.

Sep 2021 - May 2025

Indian Institute of Technology Jammu

Computer Science and Engineering

CGPA: 8.59/10

High School

April 2020 - March 2021

Army Public School, Pune

Science and Mathematics

Percentage: 96%

Projects

Transparent Surface Depth Estimation

NTIRE 2024 | March 2024

Developed a custom model to estimate depth in challenging transparent surfaces, incorporating techniques like MaskLoss and multi-loss balancing. Part of NTIRE 2024 CVPR Challenge.

Chandrayaan Moon Mapping

ISRO, IIT Tech Meet | February 2023

Applied deep learning to enhance Chandrayaan 2 imagery, achieving significant resolution boosts and optimized image handling.

BERT-based Question Answering

DevRev, IIT Tech Meet | January 2023

Led development of an optimized BERT model for rapid response times, enhancing model efficiency and accuracy.

Achievements

Positions of Responsibility