Leader profile

Marcel Torne Villasevil

Marcel Torne Villasevil is a PhD student at Stanford University. His dream is to design the next generation of learning-based assistive robots.

Current role

Role: PhD student
Company: Stanford University
Official link: https://marceltorne.github.io/

Recent viewpoints

MAM system combines short-term visual and long-term textual memory for robots.

Marcel presents MAM (multi-scale embodied memory), a system that decomposes memory into short-term visual memory for low-level dexterous control and long-term textual memory for high-level policy planning, enabling robots to track task progress and adapt to mistakes during long-horizon tasks.

Marcel Torne Villasevil Embodied memory for robot policies Source at 7:59

Recent interviews

Memory, SimToolReal, and World Action Models | YC Paper Club

Aug 8, 2026

Major topics

Design IterationLast interview Aug 15, 2026 TypographyLast interview Aug 15, 2026 InternationalizationLast interview Aug 15, 2026 Visual PrototypingLast interview Aug 15, 2026 Logo DesignLast interview Aug 15, 2026 Project PlanningLast interview Aug 15, 2026 Creative CollaborationLast interview Aug 15, 2026 Meaningful DesignLast interview Aug 15, 2026 Accessible ComputingLast interview Aug 15, 2026 Design ConstraintsLast interview Aug 15, 2026