Home
Home
I design algorithms for machine learning on time series, build scalable vector search systems, and turn research results into reliable open-source software. My work spans theory, benchmarking methodology, and end-to-end engineering.
Contact
Postal address
Humboldt-Universität zu Berlin
Department of Computer Science
Unter den Linden 6
10099 Berlin
Visiting address
Rudower Chaussee 25
12489 Berlin
Room 4.407
Phone: 030-2093-41287
Research interests
- Time series analytics: classification (BOSS, WEASEL), motif discovery (Motiflets, Leitmotifs), segmentation (ClaSP, ClaSS, ClaP)
- Vector search and high-dimensional indexing for similarity search at scale
- Benchmarking methodology and reproducible experimentation for ML systems
Teaching highlights
Recent courses (full archive on Teaching):
- WiSe 2025/26: Data Science mit Python · Vector Search (Master)
- SoSe 2025: Angewandtes Maschinelles Lernen
- WiSe 2024/25: Data Warehousing and Data Mining
- SoSe 2024: Information Retrieval · Algorithmen und Datenstrukturen
- WiSe 2023/24: Angewandtes Maschinelles Lernen · Algorithmen und Methoden der Zeitreihenanalyse
Selected publications
Recent highlights—see Publications for the full list.
- CLaP — State Detection from Time Series, PVLDB 2026.
- Fast and Exact Similarity Search in less than a Blink of an Eye, ICDE 2025.
- MotiPlus and MotiSet: Discovering the best set of Motiflets in Time Series, ECML/PKDD 2025.
- Discovering Leitmotifs in Multidimensional Time Series, PVLDB 2025.
- aeon: a Python Toolkit for Learning from Time Series, JMLR 2024.
- Raising the ClaSS of Streaming Time Series Segmentation, PVLDB 2024.
- Motiflets — Simple and Accurate Detection of Motifs in Time Series, PVLDB 2023.
- Bake off redux: A Review and Experimental Evaluation of Recent Time Series Classification Algorithms DMKD, 2024.
Service & projects
- Organizer: AALTD Workshop (2022–2025), Human Activity Segmentation Challenge @ ECML/PKDD 2023
- Program committees & reviews: KDD ’25, VLDB ’25–’26, ECML ’21–’24, AAAI ’21–’23, IJCAI ’20, TKDE, DMKD, IEEE Cybernetics, and more
- Current open-source focus: aeon — machine learning from time series
- Past projects: sktime (former core developer), MoSGrid, Harness, Contrail, XtreemFS
Open-source frameworks
- SFA/BOSS/WEASEL Java toolkit — scalable symbolic Fourier approximations
- Motiflets — Python motif discovery for time series
- Leitmotifs — subdimensional motif discovery
- aeon toolkit — Python framework for time series ML
- ClaSS — streaming segmentation
- CLaP — state detection from time series