Marie Skłodowska-Curie PhD fellowship in Learning and Optimization Algorithms for Carbon-Aware Datacenter Fleets - DTU Compute

• Designing a multi-agent orchestrator that can handle multiple datacenter sites.
• Implementing multi-timescale control, from sub-second to hourly decisions.
• Developing learning-based decision methods, such as contextual bandits with bounded-regret guarantees.
• Building a counterfactual analysis toolkit that estimates the carbon savings of different orchestration policies.
• Publishing and disseminating research in international conferences and journals.
• Collaborating with the academic and industrial partners in AEGIS, and contributing to the department with service work, including being a Teaching Assistant.

You must have a two-year master's degree (120 ECTS points) or a similar degree with an academic level equivalent to a two-year master's degree. We expect that you have:

• Strong algorithms and data structures, and solid programming skills.
• A background in one or more of: machine learning, reinforcement learning, combinatorial optimization, operations research, or distributed systems.
• Curiosity, initiative, and the drive to own a problem end to end.
• Good spoken and written English.

The standard MSCA eligibility criteria also apply:

• No doctoral degree. At the date of recruitment you must not yet hold a PhD.
• Mobility rule. You must not have lived or done your main activity (work or studies) in Denmark for more than 12 months in the 36 months just before you are recruited. Short stays such as holidays do not count.

• A letter motivating the application (cover letter)
• Curriculum vitae
• Grade transcripts and BSc/MSc diploma (in English) including official description of grading scale
• An example text written by you, such as a thesis chapter, report, paper, or substantial technical documentation
• An example code sample written by you, or a repository link with a short explanation of your contribution