Raul Sampedro

Dedicated engineer, interested in new devices and software technologies in socially impactful sectors.

San Francisco Bay Area

About

Experienced technical leader and hands-on contributor in AI, ML, autonomy, computer vision, and product development, focused on AI, robotics, and simulation with meaningful social impact. Published co-author of peer-reviewed research in reinforcement learning and robotics, contributor to a published AI book, with a track record of delivering robotics products, production ML systems and AAA games. Specialties: Technical leadership, software architecture, AI/ML and robotics systems, mentoring, rapid adaptation to new domains, continuous learning.

Experience

  • Head of Software Engineering at Sunday
    Feb 2026 - Present · 6 mos

    🤖

  • Sr. Staff Machine Learning Engineer at ICON
    Jun 2023 - Mar 2025 · 1 yr 10 mos

    • ML planning and applied research across a broad range of areas, including: Diffusion Models, Graph Neural Networks, Variational Autoencoders, 2D Image Generation and Guidance, 3D Semantic Scene Completion, 3D Reconstruction, 2D Segmentation & Classification. • Implementation of complex computational geometry solutions for Straight Skeleton Computation, Depth Estimation, Roof Generation. Key Releases: • vitruvius.ai: An AI system for designing and building homes. • https://www.cnbc.com/2024/03/19/icons-vitruvius-ai-program-designs-dream-homes.html • [open source] py_straight_skeleton library, python Straight Skeleton Computation. • https://github.com/iconbuild/py_straight_skeleton

  • Staff Software Engineer at Everyday Robots, X, Alphabet
    Jul 2022 - Apr 2023 · 10 mos

    Focused on hard-to-solve edge cases in autonomy and perception, for our robots deployed in the field.

  • Member Of Technical Staff at OpenAI
    Nov 2019 - Oct 2021 · 2 yrs

    Controls, drivers and simulation for Robotics AI Research. Systems development for Reinforcement Learning platforms. • [NeurIPS-2022] Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos • https://neurips.cc/virtual/2022/poster/54699 • Asymmetric self-play for automatic goal discovery in robotic manipulation • https://neurips.cc/virtual/2020/19903 • https://robotics-self-play.github.io/ + https://arxiv.org/abs/2101.04882 • Multi-task curriculum learning in a complex, visual, hard-exploration domain: Minecraft • https://arxiv.org/abs/2106.14876v1

  • Founding Team Software Engineer at Nines
    Dec 2017 - Nov 2019 · 2 yrs

    Core systems and applications to integrate ML into Healthcare IT environments, investor demos and production ready software. Backend, ML inference engine, front-end, and microservices. HL7, DICOM. Collaborated on establishing QMS, HIPAA compliance, IT relationships.