USC IMMERSE Lab
Graduate Research Assistant
- Built a compact JEPA-based world model pretrained on network telemetry to have one shared learned representation that can drive many network control tasks such as adaptive bitrate streaming and congestion control.
- Leading the first measurement study of WebRTC under 5G mobility using a custom Android testbed that aligns app-layer stats with radio signaling to explain mobility-induced quality drops.
- Designed a mixed reality user study that measures how latency-based cognitive attacks affect human performance in immersive fire-evacuation training.
