Soroush Bateni
Papers
1
Total Citations
88
H-Index
1
About
Soroush Bateni is a leading researcher in real-time embedded systems, with a primary focus on GPU-accelerated deep neural network (DNN) scheduling and resource management. His most cited work, "S³DNN: Supervised Streaming and Scheduling for GPU-Accelerated Real-Time DNN Workloads" (2018, 88 citations), introduces a novel framework that enables predictable and efficient execution of DNNs on GPUs for latency-critical applications like autonomous driving and robotics. This contribution directly addresses the challenge of guaranteeing timing constraints while maximizing GPU utilization—a critical bottleneck in modern embedded AI systems. Bateni’s research bridges the gap between real-time systems theory and practical DNN deployment, offering both analytical guarantees and empirical validation. His work has been recognized for its impact on safety-critical autonomous platforms, where predictable performance is non-negotiable. By developing scheduling policies that account for GPU architecture and DNN execution patterns, Bateni has helped pave the way for more reliable and responsive AI-driven systems. His ongoing research continues to shape how embedded systems handle complex, resource-intensive workloads, making him a key figure in the evolution of real-time AI acceleration.
Research Focus
Key Achievements
Top Papers
- 1