Building a chatbot with CERN

The ATLAS collaboration at CERN is one of the largest particle physics experiments in the world, and accordingly produces a tremendous amount of data. This means that resources are difficult to navigate. To overcome this, we worked on using retrieval-augmented generation (RAG) for an internal chatbot for the experiment. The idea is to help researchers and doctoral students find their way around the overload of information produced by this colossal machine.

In science, it is especially important to link results back to their original sources. This was a big emphasis of the project, to ensure that they can also access data and research papers reliably from chat queries.

A list of collaborators on the first slide of a recent presentation.

This was work I completed during a research placement for the Nuclear Center for European Research (CERN) under the supervision of Dr. Gabriel Facini.