Yoav Katz

Yoav Katz joined the IBM Research – Israel lab in 2002 where he currently manages the Language Model Utilization and Evaluation group. His team develops technologies to effectively use and evaluate LLMs, such as prompt creation, prompt optimization and model evaluation. The technologies are embedded in both IBM products and open source projects, such as Unitxt.

Before this, he managed the IBM Debating Technologies team that built Project Debater, the first AI system to successfully compete with a human in a live debate, initially joining the Project Debater team in 2014 as the chief software architect, responsible for the overall software design, infrastructure and deployment of the system.

Prior to this role, he served in various technical and management roles developing tools and methodologies for simulation-based verification of hardware designs. These technologies are used in all the IBM processors’ designs as well as by external companies.

He completed his B.Sc. and M.Sc. in computer science at the Technion Institute of Technology in Haifa, Israel. His main areas of interest are Natural Language Processing systems from both the algorithmic and engineering aspects.