Carlos E. Muñiz Cuza
AI Research Engineer · Machine Learning Systems · Time Series Compression · Time Series Analytics
I am a researcher and engineer working at the intersection of machine learning, time series compression, and systems. My work focuses on data compression, forecasting, missing value imputation, and practical open-source tools for research and deployment.
I am currently affiliated as a PostDoc with the Berlin Institute for the Foundations of Learning and Data (BIFOLD) at the Big Data Engineering (DAMS Lab) led by Prof. Matthias Boehm.
Personal Email: cmcuza@gmail.com
Work Email: muniz.cuza@tu-berlin.de
Google Scholar: profile
GitHub: github.com/cmcuza
LinkedIn: linkedin.com/in/cmcuza
Selected Projects
- TerseTS. Open-source framework unifying multiple lossless and lossy time series compression methods under a common API. [repository]
- CAMEO. A compression algorithm that preserves autocorrelation structure to maintain forecasting performance. [paper]
- SST-GCN. A graph neural network for stochastic missing-value imputation in spatio-temporal data. [paper]
Selected Publications
For a full list see Google Scholar and DBLP.
2026
- Carlos Enrique Muñiz-Cuza, Søren Kejser Jensen, Tom Louis Klein, Sabina Bakhtiiarova, Matthias Boehm, Torben Bach Pedersen: TerseTS: A Framework for Time Series Compression. EDBT 2026. [paper]
- Carlos Enrique Muñiz-Cuza, Matthias Boehm, Torben Bach Pedersen: CAMEO: Autocorrelation-Preserving Line Simplification for Lossy Time Series Compression. EDBT 2026. [paper]
2025
- Kristiyan Blagov, Carlos Enrique Muñiz-Cuza, Matthias Boehm: Fast, Parameter-free Time Series Anomaly Detection. BTW 2025. [paper]
2024
- Carlos Enrique Muñiz-Cuza, Søren Kejser Jensen, Jonas Brusokas, Nguyen Ho, Torben Bach Pedersen: Evaluating the Impact of Error-Bounded Lossy Compression on Time Series Forecasting. EDBT 2024. [paper]
2022
- Carlos Enrique Muñiz-Cuza, Nguyen Ho, Eleni Tzirita Zacharatou, Torben Bach Pedersen, Bin Yang: Spatio-temporal Graph Convolutional Network for Stochastic Traffic Speed Imputation. SIGSPATIAL/GIS 2022. [paper]
Open Source
My open-source repositories are available on GitHub.
Experience
- Research and Teaching Assistant, Technische Universität Berlin 2024-present.
- PhD Fellow, Aalborg University 2021-2024.
- Research and Development Engineer, Datys, 2015-2018.
Selected Supervision
- Master Thesis. Tom Klein (TU Berlin, 2026-present). "Optimizing Lossy Time Series Compression Pipelines for Anomaly Detection"
- Bachelor Thesis. Kristiyan Blagov (TU Berlin, 2023-2024). "Fast, Parameter-free Time Series Anomaly Detection"
- Master Thesis. Ahmed Boulila (TU Berlin 2023-2024): "Predicting the Impact of Lossy Compression on Time Series Analytics"
Teaching
- Data Integration and Large-Scale Analysis, TU Berlin (WiSe 2025/26). [course page]
Awards
- EDBT 2026 Best Demo Honorable Mention for TerseTS: A Framework for Time Series Compression.
- Erasmus Mundus Scholarship for the Joint Master's in Big Data Management and Analytics.
- Summa Cum Laude, BSc in Computer Science.
Education
- PhD in Computer Science, Technische Universität Berlin & Aalborg University, 2021-2026.
- Erasmus Mundus Joint Master's in Big Data Management and Analytics, 2018-2020.
- BSc in Computer Science, University of Oriente, Cuba. 2010-2015