Postdoc on Decision Trees for Control and Monitoring of Stochastic Systems

Our recent algorithms synthesize optimal tree policies orders of magnitude faster than earlier approaches. Now we want to take them into uncertain, real-world settings.

Postdoc on Decision Trees for Control and Monitoring of Stochastic Systems

Job description

Learned controllers are increasingly deployed in safety-critical settings, yet they offer few guarantees once reality deviates from the model they were trained on. Decision trees offer a promising alternative: they are compact, human-readable, and amenable to formal analysis.

This position builds on the recent results from our group. Possible directions include:

• Explainable monitoring: designing interpretable runtime monitors that identify which environment the system is operating in, so the right policy can be selected, or so that to detect when the deployed policy’s guarantees no longer hold.
• Scalability and stochasticity: extending search-based synthesis of decision-tree policies to stochastic and black-box systems and to larger state spaces, so that optimal tree policies remain within reach for realistic problems.

You will have the freedom to shape the direction within this scope. The aim is to publish in leading venues in formal methods and AI, along with open-source tools.

You will join the Algorithmics Group in the Department of Software Technology, Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS). You will collaborate with PhD students and postdocs working on intelligent decision-making and will have opportunities to co-supervise students. You will be supervised by Prof. Anna Lukina.

Job requirements

You should have:

• a PhD in computer science or a related field (or be close to completing one), with a background in formal methods, planning under uncertainty, or machine learning for sequential decision making;
• solid foundations in mathematics and logic, plus experience with algorithm design, combinatorial optimization, and implementation;
• a strong publication record relative to career stage, for example, at venues such as CAV, TACAS, RV, AAAI, IJCAI, or NeurIPS;
• excellent written and spoken English.

The following are a plus:

• experience with (uncertain or robust) Markov decision processes, probabilistic model checking, or reinforcement learning;
• experience with decision trees, branch-and-bound search, mixed-integer programming, or runtime verification;
• strong programming skills (e.g., Rust, Python, C++).

TU Delft (Delft University of Technology)

At TU Delft, our people make the difference. With their knowledge and curiosity, our staff provide a high-quality education and conduct pioneering research that extends beyond the campus. You will have the opportunity to take the initiative, work with others, and grow as a professional.

Faculty of Electrical Engineering, Mathematics and Computer Science

The Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS) brings together three scientific disciplines. Combined, they reinforce each other and are the driving force behind the technology we all use in our daily lives. In other words: there is plenty of room at the faculty for ground-breaking research. We educate innovative engineers and have excellent labs and facilities that underline our strong international position.

Conditions of employment

• Duration of contract is 1 year. Temporary.
• A job of 36-40 hours per week.
• Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities.
• Salary range: €3706 - €5760
• An excellent pension scheme via the ABP.
• The possibility to compile an individual employment package every year.
• Discount with health insurers on supplemental packages.
• Flexible working week.
• Every year, 232 leave hours (at 38 hours). You can also sell or buy additional leave hours via the individual choice budget.
• Plenty of opportunities for education, training and courses.
• Partially paid parental leave
• Attention for working healthy and energetically with the vitality program.

Will you need to relocate to the Netherlands for this job? TU Delft is committed to make your move as smooth as possible! The HR unit, Coming to Delft Service, offers information on their website to help you prepare your relocation. In addition, Coming to Delft Service organises events to help you settle in the Netherlands, and expand your (social) network in Delft. A Dual Career Programme is available, to support your accompanying partner with their job search in the Netherlands.

As part of knowledge security, TU Delft conducts a risk assessment during the recruitment of personnel. We do this, among other things, to prevent the unwanted transfer of sensitive knowledge and technology. The assessment is based on information provided by the candidates themselves, such as their motivation letter and CV, and takes place at the final stages of the selection process. When the outcome of the assessment is negative, the candidate will be informed.