Max Robotics — Global Service Robot Directory + US Certification

About

Doina Precup is a leading researcher in reinforcement learning (RL), whose work has fundamentally shaped how artificial agents learn, plan, and adapt across time and uncertainty. Based at McGill University and a research scientist at DeepMind, she has made seminal contributions to temporal abstraction, continual learning, and scalable AI architectures. Her early work on temporal abstraction in RL — with over 247 citations — helped establish the theoretical foundations for hierarchical decision-making, allowing agents to reason across multiple time scales. Her Horde architecture (305 citations) introduced a groundbreaking multi-agent framework enabling robots to build and maintain rich world knowledge through unsupervised interaction. Her research on bisimulation metrics and learning from limited demonstrations further broadened the mathematical and practical toolkit available to RL practitioners. More recently, Precup has championed continual reinforcement learning — the challenge of agents that learn without forgetting — contributing a widely read review (179 citations) that has become a key reference in the field. Her work on skill composition via the Option Keyboard reflects her ongoing commitment to building flexible, generalizable AI systems. Across her career, Precup's research has consistently bridged theoretical rigor with real-world applicability, cementing her as one of the most influential voices in modern machine learning.