Hi, I’m Satyam! 🙌

I’m a Master’s student in Data Science at Indiana University, Bloomington, interested in the intersection of Computer Vision, Vision Language Models, and Reinforcement Learning. Recently, I’ve been exploring post-training alignment, RLHF, and interpretability in multimodal systems, especially how reinforcement learning reshapes model behavior after pretraining.

Most recently, I worked as a Machine Learning Research Assistant at the Cyberinfrastructure for Network Science Center, collaborating with Dr. Katy Börner on a scalable computer vision pipeline for spatial biology. The work focused on cell segmentation and annotation using models like SAM and DeepCell across high-dimensional imaging formats such as OME-TIFF and spatial transcriptomics data. Before that, I worked at Jio Institute, India with Dr. Sudipta Roy on self-supervised learning for ultrasound cyst localization. My research journey originally took shape during undergraduate internships at Indian Statistical Institute under Dr. Pradipta Maji, and at K. J. Somaiya College of Engineering with Dr. Jagannath Nirmal, where I worked on marine environmental monitoring and histopathology using computer vision.

Outside research, I’m usually on the soccer field, thinking through ideas on a morning run, or looking for the kind of problem that makes me want to start a new notebook or repo. I’m a morning person, and what keeps me going is a simple reminder: Memento Mori. Time is limited, and there is still too much left to build.

You can find my full CV here and further information on my projects.

Research Interests:
                        • Mechanistic Interpretability of World Models
                        • Post-training (Reward Modeling)