I build AI systems that perceive and act in the physical world. My work spans wireless sensing using RFID, mmWave, and acoustics, GenAI for network intelligence, indoor air quality monitoring, and mobile systems research. I hold a PhD from UT Austin, have published 15+ papers at premier venues including MobiCom, CHI, SenSys, and NeurIPS, and shipped multiple AI products at Cisco.
Building multi-modal intelligent sensing systems. Using RFID, acoustic signals, and mmWave radar, we build multimodal sensing systems that track motion, measure temperature, recognize gestures, and authenticate users—enabling embodied intelligence without cameras or wearables. This is for the next generation of robotics and smart environments.
GenAI for real-world systems. Networks generate massive multimodal data—packets, logs, configs. We're building foundation models that understand this "language," enabling AI to diagnose problems, predict failures, and act autonomously. We are shipping AI products at scale, from foundational generative models for networks to agentic diagnostics.
SherlockProduct
GenAI-powered PCAP analysis for intelligent network troubleshooting
Indoor pollution is invisible but impacts health daily. From CO₂ buildup in offices to particulate matter in factories, we build systems that sense, visualize, and help people act on indoor air quality—from low-cost sensors to AR games.
COâ‚‚ ARCHI'26
AR game to find and disperse indoor COâ‚‚ hotspots
Building intelligent mobile systems. From password-free authentication using daily activities to understanding notification behaviors and optimizing mobile sensing, we explore how smartphones can become smarter, more secure, and energy-efficient through context-aware computing.
