UIUC CS PhD · InterDigital AI Lab · Wireless AI
I am Rafid Umayer Murshed, a Computer Science PhD student at the University of Illinois Urbana-Champaign and a summer ML Research Intern at the InterDigital AI Lab in Los Altos. I build neural field models, neural operators, and physics-guided learning systems for wireless propagation, sensing, and AI-native network control.
I received my B.Sc. in Electrical and Electronic Engineering from the Bangladesh University of Engineering and Technology (BUET) in May 2022, with a concentration in signal processing and wireless communications. My undergraduate thesis, under the supervision of Dr. Farhad Hossain, explored deep learning–based hybrid beamforming for MIMO systems.
After graduation, I worked as a research engineer at the Japan Institute of Disaster Prevention and Urban Safety (BUET-JIDPUS), where I contributed to the Earthquake Early Warning (EEW) project. In August 2023, I began my M.Sc. in Electrical and Computer Engineering at The University of Texas at Dallas under Dr. Mohammad Saquib, completing the degree in May 2025 with a perfect GPA of 4.0/4.0. My master’s thesis, Real-Time Beamforming and Metasurface Design for Low-Latency 6G Wireless Networks, advanced low-complexity algorithms for massive MIMO and RIS optimization.
I am currently pursuing my Ph.D. in Computer Science at the University of Illinois Urbana-Champaign (UIUC) as a recipient of the Siebel School Fellowship. I have joined the iSens Lab under Dr. Elahe Soltanaghai, where my research focuses on physics-informed deep learning for next-generation wireless systems. My work spans trustworthy digital twins, RF propagation modeling, MU-MIMO beamforming, reconfigurable intelligent surfaces, resource allocation, and wireless sensing. I have published in leading journals and conferences, and have been invited to present my research in venues such as the NSF MERIF workshop and the ARAFest. I have been fortunate to collaborate with leading professors, scholars, and researchers from premier institutions, including MIT and UC Berkeley.
WiNeRF learns continuous, complex-valued wireless channel representations from sparse commodity WiFi CSI by embedding antenna geometry, spatial resolution, and phase uncertainty as inductive biases.
Read paper
I-NAV performs questionnaire-based geolocation when GPS is unavailable, matching routing engines within minutes of travel-time accuracy while remaining offline-capable.
PU-HNO predicts high-fidelity indoor radio maps from low-fidelity ray-tracing outputs and scene priors by progressively capturing reflection, diffraction, and scattering effects. The work shows that, under conditionally unbiased label noise, the model can learn stable propagation structure and outperform noisy finite-ray training labels across image-quality and wireless deployment metrics.