Azad Singh
Postdoc Researcher, Microsoft Research India
PhD IIT Jodhpur
Hello! I am a postdoc researcher at Microsoft Research, India where I work on the CARE (Clinically Aligned Radiology Expertise) project with Mercy Ranjit.
My research interests lie in medical image analysis, with a focus on self-supervised learning, multimodal representation learning, discrete representations, and grounding clinical concepts into learned models. I work on developing interpretable and robust methods for diverse medical data modalities—including 2D imaging, 3D imaging, radiology reports, and physiological signals. Currently, I am contributing to the CARE (Clinically Aligned Radiology Expertise) project at Microsoft Research India, working towards clinically aligned radiology AI systems.
Updates
-
Jul 2026Joined Microsoft Research India as Postdoc Researcher.
-
Mar 2026Defended my PhD thesis on 16th March 2026! A heartfelt thanks to my advisor Dr. Deepak Mishra for his invaluable guidance throughout this journey. Grateful to my SRC committee—Dr. Richa Singh, Dr. Anand Mishra, and Dr. Nil Kamal Hazra—and my external examiners Dr. Suyash P. Awate (IIT Bombay) and Prof. Andreas Maier (FAU Erlangen-Nürnberg) for their time and insights.
-
Feb 2026Joined Microsoft Research India as PhD Research Intern—working on LLM-driven 3D medical segmentation.
-
2025Scale-Aware Adaptive Feature Quantization for Robust Medical Image Representation Learning accepted at ECAI (European Conference on Artificial Intelligence), Bologna, Italy. Oral presentation.
-
2025ViT Coupled Efficient CMOS Image Sensor with Sparse Acquisition and Patch Selection accepted at ISCAS (IEEE International Symposium on Circuits and Systems), London, UK.
-
2025Best Poster Runner-Up—Guide2Heart at ASMUS Workshop, MICCAI 2025, Daejeon.
-
2025OTCXR: Rethinking Self-supervised Alignment using Optimal Transport for Chest X-ray Analysis accepted at WACV (IEEE/CVF Winter Conference on Applications of Computer Vision), Tucson, Arizona.
-
2025Self-supervised Contextual Representations of Chest X-ray Images published in IEEE Signal Processing Letters.
-
2024CoBooM: Codebook Guided Bootstrapping for Medical Image Representation Learning accepted at MICCAI 2024, Morocco.
-
2024IDQCE: Instance Discrimination Learning through Quantized Contextual Embeddings for Medical Images accepted at ICPR 2024.
-
2024MLViCX: Multi-Level Variance-Covariance Exploration for Chest X-ray Self-Supervised Representation Learning published in IEEE Journal of Biomedical and Health Informatics.
-
2023Large Scale Time-Series Representation Learning via Simultaneous Low- and High-Frequency Feature Bootstrapping published in IEEE Transactions on Neural Networks and Learning Systems.
-
2022Branching Out for Better BYOL accepted at NeurIPS 2022 Workshop: Self-Supervised Learning - Theory.