DrHadi Tabatabaee

Dr Hadi Tabatabaee Malazi and his research team in the Sustainable Orchestration in Computing Continuum (SOC² Lab) study sustainable and resilient orchestration for distributed computing systems. Their work asks how AI-enabled services can be deployed, adapted, diagnosed, and recovered across edge-cloud, serverless, multi-site, and space-air-ground environments as energy use, carbon intensity, workload demand, connectivity, and failures change over time.

The lab’s research connects distributed systems, sustainable computing, resilient service orchestration, AIOps, and LLM@Edge. It is organised around four main directions.

1.        Green AI Infrastructure: Modern AI workloads increasingly run across distributed infrastructure, where energy efficiency does not always imply carbon awareness. This research investigates energy- and carbon-aware orchestration for AI workloads and digital services across edge-cloud and serverless systems. It includes resource provisioning, task scheduling, service placement, workload routing, and evaluation of trade-offs between energy efficiency, carbon impact, performance, and fairness.

2.        Self-Diagnosing Edge-Cloud Systems: Distributed edge-cloud systems are difficult to diagnose because faults propagate across services, deployments, and infrastructure layers. This research develops resilient AIOps methods for observability, telemetry abstraction, deployment-aware root-cause analysis, and recovery support in computing continuum systems. The aim is to support diagnosis and recovery under partial visibility, changing deployments, and constrained communication.

3.        LLM@Edge and Smart-Building Automation: Large language models create new opportunities for natural interaction with smart environments, but their deployment raises challenges around resource use, privacy, output correctness, and controllability. This research studies local and distributed LLMs for smart-building automation, including structured intent generation, validation of LLM outputs, serverless LLM deployment, and resource-aware execution at the edge.

4.        Resilient Computing Under Constrained Connectivity: Some distributed systems operate under intermittent connectivity, constrained resources, and changing network conditions, especially in disaster-response and space-air-ground scenarios. This research investigates service orchestration, cooperative caching, service placement, and workload adaptation for environments where connectivity and resource availability cannot be assumed stable.