Data Systems Group
Max Planck Institute for Software Systems
Building next-generation systems for massive-scale data management and AI
About the Group
The Data Systems Group (DSG) at MPI-SWS conducts cutting-edge research in systems for large-scale data management and AI. We focus on building practical, high-performance systems that address real-world challenges in data processing, machine learning, and distributed computing.
Our research spans systems for big data analytics, machine learning infrastructure, graph processing, and decentralized technologies. We collaborate with industry partners and academic institutions worldwide.
Learn more about our team
Laurent Bindschaedler
Research Group Leader
Leading research in data systems, machine learning infrastructure, and distributed computing.
Personal websiteResearch Areas
Our research spans multiple areas at the intersection of systems and machine learning
Complex AI Systems
Building agentic AI infrastructure, LLM caching frameworks, code intelligence systems, and LLM-enhanced databases.
Learned Systems
Developing learned indexes, ML-enhanced database components, and benchmarks for evaluating learned systems.
Graph Systems
Scaling GNN training, building unified graph processing frameworks, and optimizing graph partitioning strategies.
Blockchain & Web3
Researching blockchain analytics, DAO governance, decentralized prediction markets, and Web3 infrastructure.
Recent Publications
Our latest research published at top venues
Rebooting Microreboot: Architectural Support for Safe, Parallel Recovery in Microservice Systems
2026ARCS 2026
Laurent Bindschaedler
Semantic Caching for OLAP via LLM-Based Query Canonicalization
2026DOLAP 2026
Laurent Bindschaedler
SQL-of-Thought: Multi-agentic Text-to-SQL with Guided Error Correction
2025DL4C @ NeurIPS 2025
Saumya Chaturvedi, Aman Chadha, Laurent Bindschaedler
Cache Saver: A Modular Framework for Efficient, Affordable, and Reproducible LLM Inference
2025EMNLP 2025
Nearchos Potamitis, Lars Henning Klein, Bardia Mohammadi, Chongyang Xu, Attreyee Mukherjee, Niket Tandon, Laurent Bindschaedler, Akhil Arora
Latest News
Updates from the Data Systems Group
Rebooting Microreboot Paper Accepted at ARCS 2026
Paper on safe, parallel microservice recovery via a typed actuation ISA and microkernel accepted at ARCS 2026 in Mainz. Reduces agent-caused harm by 95%.
March 2026
EventHARNESS Tutorial Accepted at SOSP 2026
Half-day tutorial on hardening always-on agent runtimes, co-organized with Akhil Arora (Aarhus University). Covers isolation, tool policies, state recovery, and observability.
March 2026
PublicationSemantic Caching for OLAP Accepted at DOLAP 2026
Paper on LLM-based query canonicalization for semantic caching in OLAP systems accepted at DOLAP 2026.
February 2026
Join Our Group
We are looking for motivated PhD students, postdocs, and interns who are passionate about building systems. If you are interested in joining us, get in touch!
Get in Touch