Jan Teichert-Kluge
Research
I work on causal inference when the confounders live in text and images: double machine learning that learns from product descriptions and photos, while the causal parameter keeps valid inference.
Selected papers
-
2026
Pragmatic DML with AI-Learned Representations
arXiv preprint
When text and images enter a causal analysis as AI-learned representations, an imperfect representation biases the target by the product of two representation errors. Cross-fitted double machine learning still gives valid inference, and fold-wise fine-tuning is compatible with it.
-
2026
Valid Inference for Treatment Effects under Multimodal Confounding
Fifth Conference on Causal Learning and Reasoning (CLeaR 2026)
A neural architecture for the partially linear double machine learning model that takes text and images as confounders, evaluated on semi-synthetic data with multimodal confounding.
Earlier version DoubleMLDeep: Estimation of Causal Effects with Multimodal Data (2024)
-
2024
Adventures in Demand Analysis Using AI
arXiv preprint · cemmap working paper
Combines transformer embeddings of product text and images with tabular data to model demand for toy cars on Amazon, giving sharper sales and price predictions and more credible price-elasticity estimates.
Prices and sales in observational data are confounded. Products priced differently also differ in quality, design and positioning, and much of that information exists only in descriptions and images. Together with Victor Chernozhukov, Martin Spindler and the DoubleML team, I estimate price elasticities for e-commerce products, build neural architectures for double machine learning, and study when AI-learned representations can stand in for observed controls.
Teaching
I teach tutorials and seminars in statistics, statistical programming with Python, deep learning and causal inference at the University of Hamburg Business School, and supervise seminar papers on foundation models. Outside the university I run workshops on causal machine learning for practitioners, most recently at PyCon DE & PyData Berlin and the ZIB Academy.
Background
-
Current
Research stay, joint research project with Prof. Rahul Krishnan
Vector Institute, Toronto
-
2023–
Research associate and doctoral researcher, Statistics
University of Hamburg
-
2022–23
Data science, working student
Philips Medical Systems, Hamburg
-
2021–22
Business intelligence, working student
Siemens Gamesa, Hamburg
-
2020–23
M.Sc. Industrial Engineering
University of Hamburg / HAW Hamburg · thesis on causal machine learning with deep learning
-
2020
Hauke-Trinks-Award, 2nd place
Northern Institute of Technology Management
-
2016–20
B.Sc. Industrial Engineering
University of Hamburg / HAW Hamburg
Contact
jan.teichertkluge@uni-hamburg.de
University of Hamburg Business School, StatisticsMoorweidenstraße 18, Room 0025, 20148 Hamburg
+49 40 2395-21517