About Me

Hi! My name is Dan Graur. I’m a Senior Research Engineer in Google DeepMind where I work on Gemini Multimodality and User Signals.

I hold a PhD in Computer Science from ETH Zürich. My main research interests were in Systems for Machine Learning and Data Management.

During my PhD I interned three times in Google: (1) in Brain as part of the Flax team where I worked on a wrapper over Jax meant to ease Deep Learning research and development, (2) in TensorFlow as part of the tf.Data team where I worked on scalable and efficient ML data processing, and (3) as part of the Systems Research Group where I worked on generic, predictive, and adaptive indexing methods for databases.

Research

This is a list of the research papers I’ve published so far:

• Comanici G., …, Graur, D., …, Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities, 2025, arXiv preprint arXiv:2507.06261
• Böther M., Yao X., Kerimoglu T., Graur, D., Gsteiger V., Klimovic A., Mixtera: A Data Plane for Foundation Model Training, 2026, (preprint) Proceedings of the 2026 International Conference on Management of Data (SIGMOD)
• Graur, D., Mraz, O., Li, M., Pourghannad, S., Thekkath, C. and Klimovic, A., Pecan: Cost-Efficient ML Data Preprocessing with Automatic Transformation Ordering and Hybrid Placement, 2024, Proceedings of the USENIX Annual Technical Conference (ATC)
• Graur, D., Röthlisberger, R., Jenny, A., Drozdowski, F., Konigsmark, C., Müller, I., and Alonso, G., Addressing the Nested Data Processing Gap: JSONiq Queries on Snowflake through Snowpark, 2024, IEEE 40th International Conference on Data Engineering (ICDE)
• Audibert, A., Chen, Y., Graur, D., Klimovic, A., Simsa, J. and Thekkath, C., tf.data service: A Case for Disaggregating ML Input Data Processing, 2023, 14th Symposium on Cloud Computing
• Graur, D., Aymon, D., Kluser, D., Albrici, T., Thekkath, C. and Klimovic, A., Cachew: Machine Learning Input Data Processing as a Service, 2022, Proceedings of the USENIX Annual Technical Conference (ATC)
  • Featured in the TRC Researcher Spotlight
• [Best Paper] Graur, D., Müller I., Proffitt M., Watts G. T., and Alonso G., Evaluating Query Languages and Systems for High-Energy Physics Data, 2022, Proceedings of the VLDB Endowment
• Graur, D., Bruno, R. and Alonso, G., Specializing Generic Java Data Structures, 2021, 18th ACM International Conference on Managed Programming Languages & Runtimes
• Graur, D., Aymon, D., Thekkath, C. and Klimovic, A., Machine Learning Input Data Processing as a Service, 2021, EuroSys Doctoral Workshop 2021
• Graur, D., Bruno, R., Bischoff, J., Rieser, M., Scherr, W., Hoefler, T. and Alonso, G., Hermes: Enabling efficient large-scale simulation in MATSim, 2021, Procedia Computer Science, 184, pp.635-641
• Rellermeyer J. S., Khorasani S. O., Graur D. and Parthasarathy A., The Coming Age of Pervasive Data Processing, 2019, 18th International Symposium on Parallel and Distributed Computing (ISPDC), Amsterdam, 2019
• Graur D., Maris R. A., Potolea R., Dinsoreanu M. and Lemnaru C., Complex Localization in the Multiple Instance Learning Context, 2018, New Frontiers in Mining Complex Patterns. Springer International Publishing, Cham, 93–106

Other Contributions

I’ve also helped develop and improve the ADL Functionality Benchmarks Index, a benchmark dedicated to bridging the gap between the High-Energy Physics and the Database communities in terms of Query Languages and Database Engines:

• Proffitt M., Müller I., Graur D., Adamec M., David P., Guiraud E., and Binet S., iris-hep/adl-benchmarksindex: ADL Functionality Benchmarks Index. Version v0.1. 2021. DOI: 10.5281/zenodo.5131287

Teaching During PhD

During my time at ETH Zürich I’ve helped teach the following courses:

• Data Modeling and Databases - FS'23 & (Head TA) FS'24
• Data Management Systems - (Head TA) HS'22
• Information Retrieval - FS'22
• Big Data - HS'20 & HS'21
• Cloud Computing Architecture - (Head TA) FS'21
• Big Data for Engineers - FS'20

Contact

Feel free to get in touch by reaching out on LinkedIn or by sending me an email at hc.zhte.fni@ruarg.nad (copying the address won’t work well).