Having been recently awarded an ERC Starting grant, I will be hiring several PhD students and postdocs in the next few years. The currently active calls are listed below.

I am seeking one PhD student to work on various areas of learning theory starting from October 2023. These are fully funded full-time research positions involving no mandatory teaching duties or coursework, with an expected length of 4 years (3 years + 1 year of optional extension that is typically required for a good thesis). The successful candidates will work under my direct supervision at the AI and ML research group at UPF, with the possibility to collaborate with other researchers in the group and more broadly at UPF.

The research project broadly considers topics in learning theory, where "theory" is to be understood from the perspective of computer science and statistics. Some typical goals of theoretical work are designing learning algorithms and proving formal guarantees about their performance, or characterizing the hardness of learning problems. The present project aims to gain a deeper understanding of the generalization error of statistical learning algorithms from the perspective of information theory, convex analysis, and regret analysis. The project is a natural continuation of my recent research on generalization bounds, so glancing at my recent papers on this topic should give an idea of the nature of the work that you would be involved in.

Ideal candidates should have a strong background in mathematics in that they should be able to formalize problems and reason rigorously in a mathematical framework. Some familiarity with machine learning is expected and knowledge of theory in any related domain is a big advantage, but all of these are less important than having a good affinity for mathematics. It's important to note that formal training in math is not required, but, to put it somewhat bluntly, you are not going to have a great time if you haven't ever enjoyed proving a theorem. Good communication skills in both written and spoken English are also important. Coding skills are not essential, but some background in computer science or other engineering disciplines is useful.


Candidates should send the following material to the email address gergely.neu+phd@gmail.com with the subject line [PhD application - YOUR LAST NAME]:

• An up-to-date CV.
• Transcripts from previous studies.
• Contact details of at least two references.
• A motivation letter explaining your research interests. (Hint: It is a good idea to explain how your interests fit into the theme of this project and my research in general.)
• A sample of your technical writing. (E.g., BSc. or MSc. thesis, a preprint, or a published paper.)
• Any other documents you feel are relevant for your application.

There are also some formal requirements that are required for enrolling in the PhD program at the department. Having these ready is not necessary at the time of application, but they should be available by the starting date:

• MSc. degree in Mathematics, Computer Science, Electrical Engineering, Physics, or any related discipline.
• English language certificate (e.g., TOEFL).
  

The deadline for application is May 26 June 9, 2023. I expect to be able to sort through the applications and start interviewing shortlisted candidates by mid-June.

About the environment: Universitat Pompeu Fabra is a public university in Barcelona, Spain. It is frequently listed as one of the best universities in the country and the best young universities in the world, and is often praised for its excellent research culture and international appeal. The AI and ML group is part of the Department of Information and Communication Technologies, which is located in the pleasant Poble Nou district of Barcelona. We also maintain strong ties with the Statistics group at the Department of Economics and Business.

Please do not hesitate to reach out to me at the address gergely.neu+phd@gmail.com in case you have any questions.

I am seeking two PhD students to work on reinforcement learning theory starting from October 2021. These are fully funded full-time research positions involving no mandatory teaching duties or coursework, with an expected length of 4 years (3 years + 1 year of optional extension that is typically required for a good thesis). The successful candidates will work under my direct supervision at the AI and ML research group at UPF, with the possibility to collaborate with other researchers in the group and more broadly at UPF.

The research project broadly considers topics in reinforcement learning theory, where "theory" is to be understood from the perspective of computer science and statistics. Some typical goals of theoretical work are designing learning algorithms and proving formal guarantees about their performance, or characterizing the hardness of learning problems. The present project aims to gain a deeper understanding of the challenges posed by large-scale reinforcement learning by identifying and exploiting structural properties of Markov decision processes that make large-scale learning statistically and computationally feasible. The project is a natural continuation of my recent research on reinforcement learning, so glancing at my recent papers should give an idea of the nature of the work that you would be involved in.

Ideal candidates should have a strong background in mathematics in that they should be able to formalize problems and reason rigorously in a mathematical framework. Some familiarity with machine learning and reinforcement learning is expected and knowledge of theory in these domains is a big advantage, but all of these are less important than having a good affinity for mathematics. It's important to note that formal training in math is not required, but, to put it somewhat bluntly, you are not going to have a great time if you haven't ever enjoyed proving a theorem. Good communication skills in both written and spoken English are also important. Coding skills are not essential, but some background in computer science or other engineering disciplines is useful.

Candidates should send the following material to the email address gergely.neu+phd@gmail.com with the subject line [PhD application - YOUR LAST NAME]:

• An up-to-date CV.
• Transcripts from previous studies.
• Contact details of at least two references.
• A motivation letter explaining your research interests. (Hint: It is a good idea to explain how your interests fit into the theme of this project and my research in general.)
• A sample of your technical writing. (E.g., BSc. or MSc. thesis, a preprint, or a published paper.)
• Any other documents you feel are relevant for your application.

There are also some formal requirements that are required for enrolling in the PhD program at the department. Having these ready is not necessary at the time of application, but they should be available by the starting date:

• MSc. degree in Mathematics, Computer Science, Electrical Engineering, Physics, or any related discipline.
• English language certificate (e.g., TOEFL).
  

The deadline for application is February 12, 2021. I expect to be able to sort through the applications and start interviewing shortlisted candidates by mid-March.

Please let me know ASAP in case you have already applied to the ELLIS PhD program—I have recently been appointed as an ELLIS Scholar and will be considering applications as an advisor from this program as well.

About the environment: Universitat Pompeu Fabra is a public university in Barcelona, Spain. It is frequently listed as one of the best universities in the country and the best young universities in the world, and is often praised for its excellent research culture and international appeal. The AI and ML group is part of the Department of Information and Communication Technologies, which is located in the pleasant Poble Nou district of Barcelona. We also maintain strong ties with the Statistics group at the Department of Economics and Business.

Please do not hesitate to reach out to me at the address gergely.neu+phd@gmail.com in case you have any questions.