Saam Hosseini

Saam Hosseini

CS Ph.D. student, ML systems — checkpointing for multi-tenant LLM training. NSDI '26, SIGCOMM '26.

Publications

DeepSFU: Scalable Deepfake Detection for Video Conferencing

Published in ACM SIGCOMM, 2026

Real-time deepfake detection as a native SFU service: detects deepfake onsets directly on the encoded VP9 bitstream and offloads crypto and NIC-to-GPU frame movement to a BlueField-3 DPU — ~144× lower per-frame processing latency and 26× more concurrent conferences at 95.8% precision.

Recommended citation: T. Tran, S. Ebadi, S. M. H. Hosseini, W. Shin, E. Ram, Y. Son, S. Kim, N. Bui, K. Lee, E. Keller, S. Ha. "DeepSFU: Scalable Deepfake Detection for Video Conferencing." ACM SIGCOMM, 2026. https://dl.acm.org/doi/10.1145/3789240.3829184

eXpressSFU: Toward Super-Scalable Video Conferencing with SmartNICs

Published in USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2026

A WebRTC SFU that offloads the media data plane — SRTP crypto, packet forwarding, and simulcast/SVC layer selection — onto the Arm cores of an NVIDIA BlueField-2 SmartNIC, cutting media-packet latency 8× and serving 3× more concurrent users than the software baseline.

Recommended citation: T. Tran, S. M. H. Hosseini, S. Kim, K. Lee, N. Bui, D. Grunwald, S. Ha. "eXpressSFU: Toward Super-Scalable Video Conferencing with SmartNICs." USENIX NSDI, 2026. https://www.usenix.org/conference/nsdi26/presentation/tran