Sreevatsank Kadaveru Physical AI & Systems Researcher

I am a Ph.D. student in Electrical and Computer Engineering at UC San Diego, advised by Prof. Dinesh Bharadia in the WCSNG lab. I work on AI and Physical AI: enabling AI models to interpret multimodal sensor data, reason about the physical world, and act on it in real time.

My research is grounded in systems building. My background is in wireless systems, which informs how I approach sensing and connectivity. I have held research internships at Qualcomm, JASR Systems, and Télécom Paris, and worked as a DevOps engineer at Tata Digital. I hold an M.S. from UC San Diego and a B.Tech. in Electrical Engineering from IIT Indore.

• Physical AI
• Multimodal AI
• Large Language Models
• Robotics
• AI Systems
• Edge Computing
• Wireless Systems

Research Overview

Physical AI places demands on both models and infrastructure. A robot must interpret camera, LiDAR, and radio data, reason about its task, and act within strict deadlines, often with inference running on GPUs at the network edge rather than on the robot itself. My work addresses this challenge in three areas:

• Physical AI systems. Building RoboNEST, a testbed that connects real robots, interchangeable wireless networks, and edge GPUs, making the complete path from sensing to action observable for vision-language-action policies.
• Multimodal AI. Studying how large language models perform on RF and vision sensing tasks, isolating failures to gaps in domain knowledge, reasoning, or data interpretation, and developing architectures such as P2SIF that integrate new sensor modalities into existing LLMs without modifying their weights.
• Networks for Physical AI. Quantifying the network requirements of Physical AI applications (PAIQoO) and designing uplink schedulers for multimodal sensor flows that must be low latency, synchronized, and delivered in lock-step (mmsRAN).

Selected Publications

All publications, with BibTeX →