DC5 — Privacy-Preserving Energy Flexibility Quantification for EDCs
Host: Imperial College London, London, United Kingdom — Control and Power Research Group, Department of Electrical and Electronic Engineering.
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
Many small datacenters owned by different companies could pool their flexibility, but only if they can prove what they delivered without exposing private business data. In this project you remove that trust barrier.
You will build a way to create verifiable, fleet-wide “flexibility contracts” that do not require competing operators to hand sensitive operating data to a central party. The methods build on federated learning and reinforcement learning, with provable privacy guarantees.
What you will work on:
- Develop privacy-preserving profiling, using federated learning, to measure flexibility.
- Build uncertainty-aware flexibility models using reinforcement learning.
- Create a distributed coordination algorithm with provable privacy bounds.
What you will deliver:
- An energy-profile estimator that meets differential-privacy budgets.
- A distributed protocol that converges to good solutions.
- A report on the trade-offs between privacy and flexibility.
Secondments
You will spend two research visits with partners in the network:
- Technical University of Denmark (DTU) (Lyngby, Denmark) — 3 months from around May 2028 (academic), with Henrik Bindner: test the privacy-preserving grid-flexibility methods on the SYSLAB grid testbed.
- Shell (Netherlands) — 3 months from around May 2029 (industry), with Dan Wu: work on privacy requirements and regulation for datacenters that take part in energy markets, and define where data sharing between compute workloads and grid operators should stop.
The host
The Control and Power Research Group at Imperial College London works on the operation, planning, and economics of low-carbon power systems, with strong links to grid operators.
Supervision
Main supervisor: Fei Teng (ORCID) — Reader in Intelligent Energy Systems, Imperial College London. He works on decision-making for net-zero power systems.
Co-supervisors:
Profile we are looking for
- Machine learning (federated learning, reinforcement learning)
- Data privacy or differential privacy
- Distributed algorithms
- MSc (completed or near completion) in computer science, electrical engineering, applied mathematics, 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 DC5), and read what to include in your application.
Questions about this position? Contact the coordinator, Prof. Paul Pop, aegis@compute.dtu.dk.