PeopleofSuperintelligence

Researcher

Emir Demirović

Emir Demirović is a computer scientist at TU Delft whose work on combinatorial optimisation and constraint programming reaches into machine learning through optimal decision trees. After a PhD at TU Wien in 2017 and a postdoctoral post at the University of Melbourne, he moved to TU Delft, where he leads the Constraint Solving (ConSol) group. His MurTree algorithm, published in the Journal of Machine Learning Research in 2022, learns optimal classification trees with runtime improvements of several orders of magnitude, and his group's Pumpkin solver won MiniZinc Challenge medals in 2025 and 2026. The Association for Constraint Programming gave him its Early Career Researcher Award in 2025. He is an associate professor at TU Delft, co-directs the Explainable AI in Transportation lab, and is an ELLIS Scholar.

Timeline

  1. 2012–2017

    A PhD at TU Wien

    He places first in the ROADEF/EURO 2012 competition and earns his PhD at the Vienna University of Technology (TU Wien) in 2017 under Nysret Musliu. He also holds shorter posts at MCP, a production planning and scheduling company in Vienna, and at the National Institute of Informatics in Japan.

  2. 2017–2020

    Research fellow at Melbourne

    He is an associate lecturer and postdoctoral researcher (research fellow) at the University of Melbourne from 2017 to 2020, working with Peter J. Stuckey, James Bailey, Rao Kotagiri, Christopher Leckie and Jeffrey Chan. From 2018 on, he scores first place in the incomplete tracks of the MaxSAT Evaluation.

  3. 2019

    A predict-and-optimize paper

    In November 2019, the arXiv paper Smart Predict-and-Optimize for Hard Combinatorial Optimization Problems appears with him as second author, ahead of Peter J. Stuckey and Tias Guns. The paper investigates relaxations and warm-starting for the repeated solving step, and shows for the first time that a predict-and-optimize approach can work on large-scale combinatorial optimization problems such as weighted knapsack and complex scheduling.

  4. 2022

    The MurTree algorithm

    In 2022, the Journal of Machine Learning Research publishes MurTree, which he writes with Anna Lukina, Emmanuel Hebrard, Jeffrey Chan, James Bailey, Christopher Leckie, Kotagiri Ramamohanarao and Peter J. Stuckey. The algorithm learns optimal classification trees with dynamic programming and search, handles datasets with tens of thousands of instances and improves runtimes by several orders of magnitude over the state of the art.

  5. 2023

    Visiting the Simons Institute

    In spring 2023, he is a visiting scientist in the Extended Reunion: Satisfiability program at the Simons Institute for the Theory of Computing. He also has a spotlight paper at NeurIPS 2023.

  6. By 2024

    Leading the ConSol group

    By 2024, he is an assistant professor at TU Delft in the Netherlands. He leads the Constraint Solving (ConSol) research group, which advances combinatorial optimisation algorithms for a wide range of real-world problems, and co-directs the Explainable AI in Transportation (XAIT) lab.

  7. 2025

    ACP Early Career Researcher Award

    The Association for Constraint Programming gives him its Early Career Researcher Award in 2025. He is an invited speaker at ECAI 2025, with a talk on using search to synthesise machine learning models, control policies and certificates that people can understand, and Pumpkin, his group's constraint programming solver, takes bronze in the Fixed Search track of the MiniZinc Challenge 2025.

  8. Since December 2025

    Associate Professor at TU Delft

    He is promoted to Associate Professor at TU Delft in December 2025. In 2026, his group's Pumpkin solver earns a medal in the MiniZinc Challenge, and he is an invited lecturer on lazy clause generation at the ACP Summer School. He is an ELLIS Scholar in the ELLIS Unit Delft.

Career

  1. Since 2025Delft University of TechnologyAssistant Professor
  2. 2023Simons Institute for the Theory of ComputingVisiting Scientist
  3. 2017–2020University of MelbourneAssociate Lecturer and Postdoctoral Researcher (Research Fellow)

Education

  1. 2017TU WienPhD

From the biography.

  1. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Delft University of Technology. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Delft University of Technology. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  3. Google has the highest combined h-index among AI organizationsIncludes Delft University of Technology. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Delft University of Technology. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Emir Demirović yet. Send us a free-licensed photo.

  1. Dr. Emir Demirović emirde.github.io. Checked 21 October 2026.
  2. Emir Demirovic simons.berkeley.edu. Checked 21 October 2026.
  3. Emir Demirović tudelft.nl. Checked 21 October 2026.
  4. [1911.10092] Smart Predict-and-Optimize for Hard Combinatorial Optimization Problems arxiv.org. Checked 21 October 2026.
  5. MurTree: Optimal Decision Trees via Dynamic Programming and Search jmlr.org. Checked 21 October 2026.
  6. ELLIS Delft Talk by Emir Demirović: Optimal Decision Trees (with Constraints) via Dynamic Programming and Search tudelft.nl. Checked 21 October 2026.
  7. Early Career Researcher Award | Association for Constraint Programming a4cp.org. Checked 21 October 2026.
  8. Frontiers in AI – ECAI 2025 ecai2025.org. Checked 21 October 2026.
  9. ACP SS 2026 school.a4cp.org. Checked 21 October 2026.
  10. Emir Demirović | European Laboratory for Learning and Intelligent Systems ellis.eu. Checked 21 October 2026.