Max Robotics — Global Service Robot Directory + US Certification

About

Ksenia Konyushkova is a leading researcher in data-driven robotics and reinforcement learning, with a focus on enabling robots to learn efficiently from limited or unlabeled data. Her major contributions center on developing scalable frameworks that combine offline learning with reward sketching, allowing robots to master multiple manipulation tasks from recorded experience without costly real-world interactions. Her highly cited 2019 work, "Scaling data-driven robotics with reward sketching and batch reinforcement learning" (45 citations), introduced a practical framework that leverages large datasets of robot experience to accomplish object manipulation tasks, demonstrating how learned reward functions can replace hand-crafted ones. Konyushkova has also advanced semi-supervised reward learning for offline RL, addressing the critical challenge of training agents when interactions are expensive or unethical. Her recent work explores vision-language models as success detectors, pushing toward more generalizable reward models. With over 88 total citations across her most influential papers, Konyushkova's research is shaping the future of practical, data-efficient robot learning—making her work essential reading for anyone interested in bridging the gap between offline reinforcement learning and real-world robotic applications.