Luca Garello

Robot Planning & Autonomy Engineer

Genova, Italy
Luca Garello

Professional Summary

Robotics engineer with 6+ years of experience deploying autonomous humanoid robots in real-world, uncontrolled environments, from museum floors to hospital wards. Specialized in behavior orchestration, task and motion planning, cognitive architectures (LLM/VLM reasoning, graph memory), and full-stack ROS2/YARP engineering. Owned the end-to-end design, architectural definition, and full-stack development of the robot’s cognitive framework, driving the platform’s high-level autonomy and intelligent HRI capabilities. Strong track record shipping autonomy pipelines on physical robots with measurable societal impact.

Recent Works

Maugeri Hospital

Maugeri Hospital 2

Acquario di Genova

Building Knowledge from Interaction

Interview with iCub

Galata Museo del Mare

Experience

Postdoctoral Researcher — Head of Robot Autonomy June 2023 — Present
IIT — Contact Lab & SoftBots Lab | Genova, Italy

AlterEgo / iCub Robots — Planning, Autonomy & Cognitive Architecture

  • Designed end-to-end autonomy pipelines for live public deployments: state machines + behavior trees integrating multimodal user recognition and understanding, LLM-based dialogue (OpenAI API / local models), goal-conditioned Navigation, and safety monitoring.
  • Sole software maintainer for the full AlterEgo fleet (10+ robot units): managing the entire ROS2 codebase spanning hardware interfaces, control, navigation, and perception — covering LQR self-balancing, Nav2, VSA actuator drivers, kinematics, and Meta Quest 3S teleoperation, all migrated from ROS1 to ROS2 Lyrical via ros2_control. Extensive practical and theoretical experience in edge computing architectures such as Nvidia Jetson platforms.
  • Built graph-based RAG cognitive architecture (LangChain, LangGraph, Neo4j) enabling iCub to execute cross-user inference and aggregate interactions across different timeframes, overcoming traditional RAG memory limitations.
  • Currently investigating computationally efficient, low-parameter world models (e.g. NVIDIA's GR00T) for autonomous robotic manipulation, relying exclusively on unlabelled 2D human video as the primary observational modality.
  • Led the design and implementation of human-subject studies in real-world environments, employing co-design methodologies with key stakeholders and validated psychological scales to rigorously evaluate robot effectiveness and acceptability.
  • Field Operations & Demo Support: Managed the transport, setup, and live technical demonstrations of robots at major international conferences (including ICRA, ECCV, Humanoids and EmTech), ensuring reliable system performance in public booth environments.
Visiting Researcher March 2025
International Research Center for Neurointelligence, University of Tokyo | Tokyo, Japan
Senior Research Fellow Nov 2022 — May 2023
CONTACT Unit, Italian Institute of Technology | Genova, Italy
MATLAB & Simulink Student Ambassador Sep 2019 — Sep 2022
MathWorks | Remote
Visiting Student Researcher Mar 2019 — Jul 2022
Personal Robotics Lab, Imperial College London | London, UK

Real-World Robot Deployments

IRCCS Maugeri — ALS Clinical Ward, Milan (2026): AlterEgo operating autonomously and via teleoperation alongside clinical staff in an active ALS care unit. First humanoid robot in an Italian clinical setting. [press release]

Galata Museo del Mare, Genoa (2023–2025): AlterEgo conducting live autonomous guided tours with LLM dialogue, spatial context awareness and goal-conditioned navigation for real public visitors. [press release]

Acquario di Genova (2023–2025): AlterEgo dynamically engaging visitors to disseminate oceanographic data on the Pelagos Sanctuary, and executing intelligent, real-time query-answering regarding the marine fauna. [press release]

Stazione Marittima, Genova (2023–2025): AlterEgo inviting passengers to fill out questionnaires stimulating users to express their point of view and evaluate the services offered in the Maritime Station. [press release]

euROBIN 2024 Coopetition — 1st Place: AlterEgo successfully executed everyday household tasks using a mixed-control paradigm combining autonomous and teleoperated operation modes. [press release]

Selected Publications

Garello L., Belgiovine G., Russo G., Rea F., Sciutti A. Building Knowledge from Interactions: An LLM-Based Architecture for Adaptive Tutoring and Social Reasoning. IROS 2025.
Garello L., Cocchella F., Catalano M.G., Sciutti A., Rea F. Next-Gen Museum Guides: Autonomous Navigation and Visitor Interaction with an Agentic Robot. IROS 2026 (under eval.)
Garello L., Cocchella F., Catalano M.G., Sciutti A., Rea F. One Robot, Many Minds: Factors Shaping Visitors' Evaluation of an Autonomous Museum Robot Guide. HAI 2025.
Pasquali D., Garello L., Belgiovine G., et al. No Robot is an Island: An Always-On Cognitive Architecture for Social Context Awareness. IEEE ICDL 2025.

Education

PhD in Cognitive Robotics 2019 — 2023
IIT + MaLGa + University of Genova

Thesis: Learning Actions from Human and Robot Visual Motion Observations
Advisors: A. Sciutti, F. Rea, N. Noceti

M.Sc. in Robotics Engineering 2018 — 2019
University of Genova
B.Sc. in Biomedical Engineering 2016 — 2018
University of Genova

Technical Skills

Robotics: ROS2 Jazzy/Lyrical, ros2_control, Nav2, YARP, MoveIt2, Gazebo, MuJoCo
Planning & Autonomy: Behavior trees, state machines, task planning, motion planning, LQR, sim-to-real
AI / LLM: LangChain, LangGraph, OpenAI API, LLM tool calling, VLM/VLA, graph-based RAG, Neo4j
Perception: Face recognition/Identification, ASR/TTS, depth cameras, point cloud (PCL), OpenCV
Hardware: Nvidia Jetson, VSA actuators, IMU, F/T sensors, Meta Quest 3S, Unity ROS2 bridge
Programming: Python (advanced), C/C++ (intermediate), MATLAB, PyTorch, Tensorflow, Docker, Git

Languages

Italian (native)  |  English (professional)  |  Spanish (intermediate)