Harald Gustafsson

Papers

Total Citations

H-Index

About

Harald Gustafsson is a rising researcher at the forefront of edge computing and real-time systems, specializing in the intersection of video analytics, resource management, and fault tolerance. His work addresses the critical challenge of delivering precise, low-latency processing for live video streams from mobile cameras—such as those on cars or robots—where fluctuating GPU loads on shared edge infrastructure demand intelligent resource orchestration. His most cited work, "RAVAS" (2023, 6 citations), introduces an interference-aware framework for model selection and resource allocation in live edge video analytics, pioneering adaptive strategies that balance accuracy and latency under dynamic conditions. More recently, with "RTilience" (2025, 2 citations), Gustafsson tackles the complex problem of configuring fault-tolerant, time-critical service chains with arbitrary DAG topologies, prototyping a scalable solution on Kubernetes. This work extends cloud-native principles to safety-critical edge applications, ensuring resilience without sacrificing timing guarantees. Though early in his career, Gustafsson’s contributions are already shaping how next-generation edge systems handle the convergence of mobility, real-time constraints, and reliability—a vital area for autonomous systems and smart infrastructure.