Katarzyna Reluga

Katarzyna Reluga

Assistant Professor

Humboldt University of Berlin

Biography

I am an Assistant Professor in AI in Business and Economics in the School of Business and Economics at the Humboldt University of Berlin. Prior to joining Humboldt University, I was a Lecturer in Statistical Science in the School of Mathematics at the University of Bristol. I also held postdoctoral positions at the University of California, Berkeley, the University of Toronto, and the University of Cambridge, where I worked with Mark van der Laan, Dehan Kong, and Qingyuan Zhao, respectively. I completed my PhD in the University of Geneva, where I was advised by Stefan Sperlich and co-advised by María José Lombardía.

I develop statistical and machine learning methods for reliable inference and prediction from scarce, incomplete, or heterogeneous data. My research connects survey methodology, causal inference, missing data analysis and machine learning. I look for parallels in the assumptions and methods used across these fields and build on them to develop tools suited to the data available in practice. I apply these methods to problems in official statistics, public health, and policymaking.

Pronouns: she/her

Nickname: Kasia

Interests
  • Causal Inference
  • Machine Learning
  • Mixed and Multilevel Models
  • Resampling Techniques
  • Survey Statistics
Education
  • Ph.D. in Statistics, 2020

    University of Geneva

  • M.S. in Statistics, 2016

    KU Leuven

  • Exchange Student, 2016

    ETH Zurich

  • B.S. in Quantitative Methods in Economics and Information Systems, 2014

    Warsaw School of Economics

  • B.A. in Modern Languages, 2014

    University of Warsaw

Publications & Preprints

(2026). Small area learning. arXiv preprint (submitted to the Annals of Statistics).

PDF Cite Source Document

(2026). Direct Doubly Robust Estimation of Conditional Quantile Contrasts. ICLR 2026.

PDF Cite Source Document

(2025). The impact of job stability on monetary poverty in Italy -- causal small area estimation. Journal of the American Statistical Association.

PDF Cite Code

(2024). Conditional Outcome Equivalence A Quantile Alternative to CATE. NeurIPS2024.

PDF Cite

(2024). Bootstrap-based statistical inference for linear mixed effects under misspecifications. Computational Statistics and Data Analysis.

PDF Cite Code

(2024). A unified analysis of regression adjustment in randomized experiments. Electronic Journal of Statistics.

PDF Cite

(2022). Simultaneous inference for linear mixed model parameters with an application to small area estimation. International Statistical Review.

PDF Cite

(2021). Simultaneous inference for empirical best predictors with a poverty study in small areas. Journal of the American Statistical Association.

PDF Cite

Contact

  • katarzyna.reluga@hu-berlin.de
  • Humboldt-Universität zu Berlin, School of Business and Economics, Chair of AI in Business and Economics, Unter den Linden 6, Berlin, 10099