Souraja Kundu

I am currently a Machine Learning Ph.D. student at Carnegie Mellon University. I am fortunate to be supervised by Prof. Laszlo A. Jeni. My research interests are grounded reasoning and post-training of multimodal models. I am currently interning at NVIDIA under Ratnesh Kumar in the data-enhanced VLM fine-tuning team. I am also collaborating with the Amazon AGI team on video understanding and VLM reasoning. Besides, I am part of a project in collaboration with Apple on understanding the spatial reasoning capability of modern MLLMs based on video demonstrations. In July 2025, I graduated from the Indian Institute of Technology Guwahati with the Governor of Assam Gold Medal. I have also explored research in medical imaging, satellite imaging, vision-audio models, and efficient machine learning architectures. I spent my post-sophomore summer in Japan working at Iwahori Vision Lab. I also worked as an R&D intern at Samsung Research Institute in 2024. I am fortunate to work with Prof. Yuji Iwahori at Chubu University, Japan, Prof. Gaurav Trivedi, Dr. Manish Bhatt, and Prof. M.K. Bhuyan at IIT Guwahati, Prof. Sachin B. Patkar at IIT Bombay, Karl F. Macdorman at Indiana University, and Dr. Farhad Merchant at the University of Groningen, Netherlands. I am humbled to receive the Quad Fellowship '25-'26, IndiaAI Fellowship '24-'25 and Institute Merit Scholarship (IIT Guwahati) '23-'24 and '24-'25. 

Fun Facts: 

• My Erdős number is 4, and my co-authorship distance from Geoffrey E. Hinton is 4!

• My motivation in Computer Vision is for Astronomical Image Processing and, in general, learning about outer space: My first writing sample.