TU Wien Informatics

Andrija Stanišić

Andrija Stanišić

Research Scientist and PhD Student

BSc, MSc

+43 1 58801 18427

LinkedIn LinkedIn Google Scholar Google Scholar ORCID ORCID

Short CV

Andrija Stanišić is a PhD student and Project Assistant at the Distributed Systems Group at TU Wien. His research focuses on engineering distributed artificial intelligence (AI) systems across the edge-cloud-space continuum. In particular, he develops methods for managing distributed training and inference under heterogeneous resources, intermittent connectivity, and limited and time-varying energy budgets. His work investigates how workload configuration, placement, partitioning, routing, and participant selection can balance energy consumption against service-level objectives.

He received his MSc and BSc degrees in Electrical and Computer Engineering from the Faculty of Technical Sciences at the University of Novi Sad. He completed his MSc in 2024 with a specialization in High-Performance Computing. In January 2025, he joined the Distributed Systems Group at TU Wien to pursue his PhD.

Research Interests

  1. Distributed AI Systems. Design and operation of distributed AI workloads across heterogeneous computing infrastructure. The research considers the interdependent system decisions required to satisfy application-level performance objectives.

  2. Distributed Training and Inference Optimization. Methods for configuring, placing, partitioning, and coordinating distributed training and inference workloads. This includes workload placement, inference routing, training configuration, participant selection, and training-informed adaptation of AI models at runtime.

  3. Energy- and Performance-Aware Optimization. Models and optimization mechanisms that balance energy consumption against service-level objectives such as latency, throughput, prediction quality, and training progress. Particular attention is given to systems with limited and time-varying energy budgets.

  4. Edge-Cloud-Space Continuum. Execution and orchestration of AI workloads across cloud data centers, terrestrial edge infrastructure, airborne platforms, and orbital computing systems. The research addresses heterogeneous resources, intermittent connectivity, dynamic network topologies, and constrained computing capacity.

Thesis Co-supervision

Students interested in pursuing a bachelor’s or master’s thesis in these or related research areas are welcome to get in touch.

Master’s Students

  1. Jaime Gallego Chillón. Design and Development of a Predictive Energy-Aware Client Selection Strategy for Federated Learning in LEO Constellations. (completed)

  2. Maximilian Maresch. Distributed Training of Deep Learning Models in the Edge-Cloud-Space Continuum. (completed)

Projects

NexaSphere

NexaSphere is a Horizon Europe research and innovation project focused on designing a next-generation three-dimensional network of networks for future 6G systems. It integrates terrestrial, airborne, and spaceborne infrastructure to enable seamless multi-connectivity and improve network availability for applications such as future air mobility, smart cities, and mobile transportation systems.

As a Research Scientist in NexaSphere, my work focuses on energy-efficient distributed inference and model training, Compound AI systems for resource management across the edge-cloud-space continuum, and serverless execution models for distributed inference.

Funding. European Union Horizon Europe Research and Innovation Programme, Grant Agreement No. 101192912
Website. nexasphere.eu

Publications

  1. M. Gravara, A. Stanisic, and S. Nastic. “Atlas: Optimizing Deployment of Compound AI Workflows on Heterogeneous Clusters.” In Proceedings of the IEEE/ACM Symposium on Edge Computing, SEC (2026) (paper).

  2. A. Stanisic, M. Gravara, J. L. Herrera, and S. Nastic. “Constella: A Novel Framework for Cost-Efficient Distributed AI Inference in LEO Space Data Centers.” In Euro-Par: Parallel Processing, 32nd European Conference on Parallel and Distributed Processing, Proceedings, Part II, pp. 347–362 (2026) (paper).

  3. M. Gravara, A. Stanisic, and S. Nastic. “Design Methodology and Performance Trade-offs Management for Distributed and Compound AI Systems.” In Proceedings of the IEEE International Conference on Cloud Computing, IEEE CLOUD (2026) (paper).

  4. M. Gravara, C. Marcelino, A. Stanisic, and S. Nastic. “PLAIground: SLO-Driven Runtime Model Selection for Compound AI Systems in the Edge-Cloud-Space Continuum.” In Proceedings of the IEEE International Conference on Smart Computing Workshops and Other Affiliated Events, SmartComp Companion, pp. 261–266 (2026) (paper).

  5. M. Maresch, A. Stanisic, and S. Nastic. “LoftNN: Distributed Training of Deep Learning Models in the Edge-Cloud-Space Continuum.” In Proceedings of the 7th IEEE International Conference on Autonomic Computing and Self-Organizing Systems, ACSOS (2026).

  6. M. C. Kaya, T. W. Pusztai, A. Stanisic, and S. Nastic. “Currus: A Compound AI Approach to Distributed Vehicle Trajectory Reconstruction in the Edge-Cloud.” In Proceedings of the 15th International Conference on the Internet of Things, IoT, pp. 254–262 (2025) (paper).

  7. A. Stanisic and S. Nastic. “ProbSelect: Stochastic Client Selection for GPU-Accelerated Compute Devices in the 3D Continuum.” In Proceedings of the 3rd International Conference on Federated Learning Technologies and Applications, FLTA (2025) (paper).

  8. M. Gravara, A. Stanisic, and S. Nastic. “A Novel Compound AI Model for 6G Networks in 3D Continuum.” In Proceedings of the European Conference on Networks and Communications & 6G Summit, EuCNC & 6G Summit (2025) (paper).

Awards

  1. Best Artifact Award. Awarded by Euro-Par 2026, the 32nd International European Conference on Parallel and Distributed Computing, in August 2026 for the artifact accompanying “Constella: A Novel Framework for Cost-Efficient Distributed AI Inference in LEO Space Data Centers.” The award recognized the originality, technical excellence, and impact of the artifact developed at TU Wien (artifact).