I am a computer vision researcher and biomedical engineer with expertise in developing multimodal models and finetuning foundation models for on-edge deployment on robots, and in clinical settings. I am currently a core ML team member at the DARPA Triage Challenge, developing multimodal deep learning models for battlefield trauma detection. The bulk of my work at CMU has centered around this role - rapid testing and deployment of algorithms for physiological sign sensing using robots. Given our aggressive deployment-first approach, I have more cool videos to show than NeurIPS publications - see them here.
Aside from my work on multimodal algorithms for medical triage robotics, I have worked on the below topics:
- Machine Unlearning for privacy-preserving AI
- Imitation Learning for long-horizon task planning in embodied AI simulations
- Uncertainty-aware segmentation models for tumor and pleural effusion segmentation
📌 Check out my Projects!!
✍️ For my thoughts on AI & research, visit the Blog. 📄 My resume (PDF)
