Core Team
Kilian Bänziger
Next-Gen RAG
Master StudentAbout
Kilian Bänziger is a statistician and data scientist working on context engineering for retrieval augmented generation pipelines. His research contributes to improve reasoning in the frame of organizational understanding to enable agents to successfully accomplish tasks across organizational context. Before joining ASL he worked on time series forecasting software at PwC and as a forward deployed engineer at the procurement startup Tacto.
Research Areas
Connect
Project
MetaRAG: A metadata network to assemble the context puzzle
MetaRAG is an approach to combine GraphRAG and Vector RAG to address the tradeoff between reasoning quality and scalability on large amounts of data in organizational contexts. The novel RAG architecture aims to reduce the number of edges and vertices in a knowledge graph to its most essential elements, which can be inferred from metadata. This enables agents to reason across large amounts of information without running into the challenges of knowledge graph creation and exponentially growing memory footprints.
Other team members
Students
Interested in collaborating?
We are always looking for talented students, researchers and industry partners.
