DC9 — Intent-Driven Power Management and Facilities Control

Host: Lund University, Lund, Sweden — Department of Automatic Control (LTH).
Work package: WP2 (Carbon-Aware Intelligence and Control)
Duration and start: 36 months, full-time; start between February and September 2027 (latest start September 2027).

The project

There is a gap between a high-level goal like “cut carbon now” and the low-level, vendor-specific controls of a site’s cooling, batteries, and compute hardware. In this project you close that gap.

You will build a control layer that takes “intents” and translates them into a coordinated set of actions across the site’s physical assets. The models tune themselves online using reinforcement learning.

What you will work on:

• Design an intent language that captures carbon, flexibility, and cost.
• Break high-level intents down into commands for the physical assets.
• Tune the power-performance and thermal models online using reinforcement learning.

What you will deliver:

• Control software that turns intents into commands, with measured latency.
• A coordinated control library with documented power ramp rates from experiments.
• A telemetry system that reports headroom and runs without breaking service-level limits in test scenarios.

Secondments

You will spend two research visits with partners in the network:

• École Polytechnique Fédérale de Lausanne (EPFL) (Lausanne, Switzerland) — 3 months from around January 2028 (academic), with David Atienza: build and test the site controller on EPFL’s EcoCloud.
• Elisa (Finland) — 3 months from around September 2028 (industry), with Simon Holmbacka: test the intent-handling interface for compute and facilities on Elisa’s aggregation platform.

The host

The Department of Automatic Control at Lund University (LTH) is a world leader in control engineering, system identification, and real-time computing.

Supervision

Main supervisor: Johan Eker (ORCID) — Professor at the Department of Automatic Control, Lund University (shared position with Ericsson); leads the WASP WARA-Ops arena. He works on cloud computing and control.

Co-supervisors:

• David Atienza (ORCID) — EPFL
• Simon Holmbacka — Elisa

Profile we are looking for

• Control engineering and/or reinforcement learning
• Systems software
• Knowledge of energy systems or building/facilities control is a plus
• MSc (completed or near completion) in control engineering, computer science, electrical engineering, or a related field
• Curiosity, initiative, and good spoken and written English

You must also meet the MSCA eligibility rules — no PhD yet, the mobility rule, and open to all nationalities.

How to apply

Official vacancy at the host institution: link coming soon.

Until then, apply through the central AEGIS application portal: AEGIS application portal (reference position DC9), and read what to include in your application.

Questions about this position? Contact the coordinator, Prof. Paul Pop, aegis@compute.dtu.dk.