About Me
My name is Sergi Sanchez Orvay, and I am an aspiring robotics researcher dedicated to advancing the field of embodied intelligence. I am currently a Robotics Research Intern at BMW Group’s RoboTac Lab in Munich, working under the supervision of Prof. Dr. Mohsen Kaboli, while pursuing my MSc in Robotics, Systems and Control at ETH Zürich. My academic foundation was built at Universitat Politècnica de Catalunya, where I completed my BSc in Electronic Engineering and Telecommunications and conducted research at the IRI-CSIC (Institut de Robòtica i Informàtica Industrial) on vision algorithms for event-cameras.
My work and interests are focused on pioneering the next generation of intelligent, general-purpose robots. My approach in embodied AI and robotic foundation models emphasizes leveraging representation learning and world models to create systems that learn from interaction and generalize across complex robotic manipulation tasks. My ultimate goal is to drive the efficiency of this learning process, bringing it closer to how humans acquire knowledge about the world through experience.
Experience
Robotics Research Intern, BMW Group (RoboTac Lab)
Munich, Germany | Sept. 2025 – Present
Working at the RoboTac Lab in BMW Group under the supervision of Prof. Dr. Mohsen Kaboli focusing on Embodied Intelligence.
- Conducting research at the intersection of embodied intelligence, cross-modal perception and robotic manipulation.
- Building robotic foundation models that unify tactile sensing, visual observations and action sequences into a shared representation.
- Contributing to the PHASTRAC project (Horizon Europe), utilizing Oscillatory Neural Networks to advance efficient edge intelligence for robotics.
Student Researcher, ETH Zürich (Vision for Robotics Lab)
Zürich, Switzerland | Feb. 2025 – July 2025
Working on Gaussian Belief Propagation (GBP) for Continuous-Time SLAM at the Vision for Robotics Lab (V4RL) supervised by Xinyi Li, William Talbot, Dr. David Hug, Dr. Cornelius von Einem and Prof. Dr. Margarita Chli.
- Focused on analyzing and resolving the numerical instabilities of Hyperion (the first open‑source continuous‑time GBP solver) that prevented its deployability.
- Designed and implemented improved regularization strategies, including diagonal and Levenberg‑Marquardt regularization, message damping and spline tail‑fixing, enabling stable online optimization.
- Improvements yielded millimeter‑level trajectory accuracy, consistent convergence across synthetic and indoor sequences, and runtime reductions of more than 100× compared to Ceres‑based Non‑Linear‑Least‑Squares solvers, demonstrating the viability of GBP as an efficient continuous‑time state estimation framework.
Robotics Research Intern, Institut de Robòtica i Informàtica Industrial (IRI-CSIC)
Barcelona, Spain | Sept. 2023 – July 2024
Research on lightweight, interpretable algorithms for motion and feature estimation using event cameras in resource-constrained aerial robotic platforms performing agile motion (accelerations of up to 40 rad/s² and 15 m/s²). Worked at the Mobile Robotics Lab under the supervision of Dr. Juan Andrade Cetto.
- Developed and implemented methods for event‑based optical flow estimation using plane‑fitting techniques on surfaces of active events.
- Implemented a state‑of‑the‑art robust corner detection algorithm in Python.
- Designed a geometric model‑based pipeline for egomotion estimation, combining normal flow, inverse depth optimization and RANSAC‑based motion recovery for my Bachelor’s Thesis.
Education
MSc in Robotics, Systems and Control
ETH Zürich | Sept. 2024 – Present
Pursuing advanced training in robotics with a focus on embodied intelligence and robot learning & perception under the academic supervision of Prof. Dr. Roland Siegwart.
- Strong theoretical foundation in deep learning, robot control, probabilistic AI and computer vision, combined with extensive hands‑on project experience.
- Semester Project: Gaussian Belief Propagation for Continuous‑Time SLAM with the Vision for Robotics Lab.
- Research Project: Towards Depth‑Guided Self‑Supervised World Models with the Computer Vision and Geometry Group.
- Internship at BMW Group (RoboTac Lab) conducting research in cross‑modal perception and tactile intelligence.
BSc in Electronic Engineering & Telecommunications
Universitat Politècnica de Catalunya | Sept. 2020 – July 2024
Obtained fundamental foundations in mathematics, physics, information theory, electrical engineering and control theory.
- Ranked 1st of the class (GPA: 9/10) receiving the Best Academic Transcript Award.
- Honors in +30% of the courses.
- Research internship at IRI‑CSIC working on vision algorithms for event‑cameras that lead to my bachelor’s thesis: Event‑based egomotion estimation (Grade: 9.8/10).
Awards & Scholarships
- Nova Talent Member (Professional network connecting and mentoring top talent worldwide) | Nova - Nov. 2024
- JAE Intro ICU - Robotics Research Scholarship | Consejo Superior de Investigaciones Científicas (CSIC) - 2023
- Finalist in the Rohde & Schwarz International Engineering Competition | Rohde & Schwarz - 2023
- Best Student Award in the Initial Phase of Electronic Eng. & Telecommunications degree | UPC - 2021
- Academic Excellence Award (Full financial support for the first university year) | Govern de les Illes Balears - 2020