Moshe Mash, Ph.D.

I am an AI researcher who studies how people and intelligent agents can work well together. My work brings together multi-agent systems, game theory, and machine learning, with an applied line of research in natural language processing and large language models.

My research follows three main directions. The first asks how humans and computer agents interact, coordinate, and reach agreements in shared environments, and how to design agents that act well alongside people. The second studies the people who build AI: how machine learning practitioners and data scientists work, and where they get stuck, so that tools can help at the right moment. The third develops LLM-based methods that extract structure and meaning from large collections of specialized text, for example predicting the authors of historical texts whose attribution has long been disputed.

I hold a Ph.D. in Software and Information Systems Engineering from Ben-Gurion University and was a postdoctoral fellow at Carnegie Mellon University. My work received the Best Paper Award at the ACM Conference on Economics and Computation (EC 2016) and was selected as a Communications of the ACM (CACM) Research Highlight (2018), with publications in JACM, AIJ, ACM TIST, CHI, IJCAI, and AAMAS.

Alongside research, I have more than a decade of hands-on industry experience taking AI systems from research to production. At Amdocs I designed and deployed, end to end, a churn-prediction system that saved millions of dollars; at DICTA I develop AI models for understanding large collections of text; and through MashInnovateAI, the AI R&D consultancy I founded, I help companies solve their core challenges with AI.

I have taught at five institutions, twice received Ben-Gurion University's Excellence in Teaching Award, and mentored students at Ben-Gurion University and Carnegie Mellon University. I serve as a reviewer for AAAI, IJCAI, AAMAS, KDD, CHI, EC, and ECAI.