About Publications Teaching CV

UPC · BarcelonaTech

Dimitris Selis

PhD Candidate · Cloud-native Systems · Elastic resource management · ML for Systems

Cloud–edge
Distributed systems research
ML / DRL
Systems reliability & performance
Kubernetes
SLA-aware orchestration

Research profile

PhD work at UPC on cloud–edge systems, learning-based control, and reliability in cloud-native environments.

01

About

I am a PhD candidate at the Universitat Politècnica de Catalunya (UPC) working on applied research in cloud–edge distributed systems, with a focus on learning-driven scheduling, autoscaling and reliability under highly volatile workloads. My research bridges large-scale infrastructure with learning-based control and robust optimization, aiming to improve service availability, trustworthiness and performance under uncertainty in cloud-native environments.

02

Research scope

Cloud-native · Learning-based control · Reliability

  • Cloud-edge infrastructure under uncertainty
  • K8s-native orchestration for elastic, SLA-aware services
  • ML / DRL for systems reliability and performance
  • Large-scale distributed systems
  • Reliability & performance optimization
04

Research highlights

All publications →

MIND the Proximity

Transformer-based DRL scheduler for service provisioning in cloud-edge systems.

↑ SLASatisfy
↓ P95Latency
StableScale

IEEE TCC, under review

KARMA

Zero-touch DRL autoscaling for Kubernetes edge workloads beyond HPA.

↓ SLAViol.
↓ P95Latency
StableScale

IEEE GLOBECOM 2025 · Xplore

ADDAPT6G

Dueling DDQN for service placement in LXD containers.

↓ SLAViol.
RobustPolicy
OODStable

IEEE ICC 2025 · Xplore