This website — just like everything else — is Work in Progress

Matteo Macchini

Matteo Macchini

About Me

Hello stranger! I am Matteo — engineer, AI researcher, recovering roboticist, and body language nerd. I build end-to-end machine learning systems that understand and model human behavior — from sensor data and model training to on-device deployment — spanning robotics, human-machine interaction, and digital humans.

My path started at CERN, designing motion control systems for particle beam diagnostics. I then spent four years at EPFL researching body-machine interfaces for robot teleoperation — work cited over 200 times and recognised with the Best PhD Thesis award in 2021. Now I lead human understanding and avatar systems at Magic Leap, bringing real-time human presence into mixed reality.

Current Role

Staff ML/AI Engineer — Human Understanding & Digital Humans

Magic Leap — Zurich, CH

Education

PhD in Machine Learning for Robotics, with Distinction

EPFL · Lausanne, 2021

Award

Best PhD Thesis in Robotics

EPFL · 2021

Institution ranking in Robotics

#1 in Europe#8 worldwide

Skills & Tools

AI & Machine Learning

Machine LearningDeep LearningMultimodal AIPyTorchCUDA

Vision & Perception

Computer VisionBody TrackingPose EstimationAction RecognitionSensor Fusion

XR & Digital Humans

XR / ARAndroid XRDigital HumansUnityOpenXR

Languages

PythonC++C#

Languages

Italian

Native

English

Fluent

French

Fluent

German

Basic

Let's be in touch!

Experience

Staff ML/AI Engineer — Digital Humans

Sep 2024 — Present

Magic Leap · Zurich, CH

Embedded collaboration (TVC) with a leading technology partner on XR initiatives

  • Leading AI/XR initiatives focused on avatars, embodied agents, and body tracking for XR headsets and glasses.
  • Shipped prototype systems integrating perception, Gemini, and full-body tracking, including an official public Android XR Unity sample showcasing avatars.
  • Driving research and production efforts in pose estimation, sensor fusion, and generative AI.

Staff ML/AI Engineer — Digital Humans

Sep 2021 — Present

Magic Leap · Zurich, CH

  • Led a 5-engineer team to deliver the core ML2 Avatars system for collaborative XR experiences.
  • Contributed to defining the strategic direction and roadmap for the Avatars initiative with leadership.
  • Led end-to-end development of AI/ML systems including hand pose classification and action recognition.
  • Productized key AI/ML technologies into company-wide platform capabilities for ML2.
  • Owned the Avatar software stack across Unity/C#, networking, animation, and data collection pipelines.
  • Mentored interns and drove alignment across research, engineering, product, and design teams.

Doctoral Researcher — ML for Human-Robot Interaction

Jul 2017 — Sep 2021

EPFL · Lausanne, CH

  • Developed state-of-the-art machine learning methods for body motion data processing.
  • Published 7 first-author peer-reviewed papers in top-tier journals and international conferences.
  • Teaching assistant in 6 master courses including aerial and mobile robotics, machine learning.
  • Supervised 25+ student projects and coordinated multi-institution research collaborations.
  • Co-authored 2 successful grant proposals in collaboration with international academic and industrial partners.
  • Initiated 3 international collaborations with top universities across Europe.

Data Analyst & Lab Coordinator

Jul 2015 — Sep 2017

CERN · Geneva, CH

  • Led development of an automated data analysis software for superconducting wire diagnostics.
  • Responsible for experiment scheduling based on data analysis.
  • Increased laboratory efficiency by introducing automated experiments and improved diagnostics.
  • Commissioned, tested, and installed 6 industrial furnaces.
  • Designed and implemented the corresponding acquisition and control system in LabVIEW.

Publications

  1. 2025Efficient Egocentric Action Recognition with Multimodal Data. EgoVis @ CVPR 2025
    M. Calzavara, A. Kastrati, M. Macchini, D. Vasilevski, R. Wattenhofer
    Full paper in preparation
  2. -Online Adaptation of Body-Machine Interfaces Based on Classification of Body Motion Patterns. IEEE Trans. SMC · In Review
    M. Macchini, Y. Belal, A. Giusti, G. Abbate, M. Tognon, F. Schiano, R. Siegwart, and D. Floreano
  3. 2024Data-Driven Personalization of Body-Machine Interfaces to Control Diverse Robot Types. IEEE Trans. SMC
    M. Macchini, F. Schiano, and D. Floreano
  4. 2021Does spontaneous motion lead to intuitive Body-Machine Interfaces? A fitness study of different body segments for wearable telerobotics. IEEE RO-MAN
    M. Macchini, J. Frogg, F. Schiano, and D. Floreano
  5. Arm-Wrist Haptic Sleeve for Drone Teleoperation. IEEE RA-L
    V. Ramachandran, M. Macchini, and D. Floreano
  6. Personalized Human-Swarm Interaction through Hand Motion. IEEE RA-L
    M. Macchini, L. De Matteïs, F. Schiano, and D. Floreano
  7. The Impact of Virtual Reality and Viewpoints in Body Motion Based Drone Teleoperation. IEEE VR
    M. Macchini, M. Lortkipanidze, F. Schiano, and D. Floreano
  8. 2020Hand-worn Haptic Interface for Drone Teleoperation. IEEE ICRA
    M. Macchini, T. Havy, A. Weber, F. Schiano, and D. Floreano
  9. 2019Personalized Telerobotics by Fast Machine Learning of Body-Machine Interfaces. IEEE RA-L
    M. Macchini, F. Schiano, and D. Floreano
  10. 2017Design and validation methodology of the control system for a particle beam size measurement instrument at the CERN laboratory. ACC
    J. Emery, A. Barjau, B. Dehning, J. Herranz Alvarez, P. Lapray, and M. Macchini
  11. 2016The bundle-barrier PIT wire developed for the HiLumi LHC project. IEEE Trans. Appl. Supercond.
    B. Bordini, A. Ballarino, M. Macchini, D. Richter, B. Sailer, M. Thoener, and K. Schlenga

Oral Presentations & Invited Talks

  • Open Science in Practice Body Motion as data source for Telerobotics. (Lausanne, 2017)
  • Humans in Motion 2018 Drone piloting through Body Motion. (Heidelberg, 2018)
  • ICRA 2020 Hand-worn Haptic Interface for Drone Teleoperation. (Paris — online, 2020)
  • IEEE VR 2021 The Impact of Virtual Reality and Viewpoints in Body Motion Based Drone Teleoperation. (Lisbon — online, 2021)
  • IEEE Neuro-EMBS 2021 Personalized Body-Machine Interfaces for Advanced Human-Robot Interaction. (Online, 2021)