F.B. / Manifesto
— 01 — MMXXVI
Robotics · Perception · AI

Federico   Bennasciutti

Robotics & AI Engineer / Leuven · BE

Teaching machines to see the world before they move through it.

— manifesto —

Machines that understand before they act.

Most of what makes a robot useful is invisible: the part where it figures out the world before it does anything. That’s the part I work on. The goal is technology that ages well — quiet, careful, and built to be trusted.

01 Life2016 — present
2022 → Leuven, Belgium

Senior Robotics Engineer // Intermodalics

  • Building perception and SLAM pipelines for autonomous mobile robots, integrating 2D/3D cameras, LiDAR, and GNSS.
  • Designing and optimising AI detection and tracking stacks for NVIDIA Jetson edge platforms.
  • Built a near-realtime navigation stack for autonomous logistics robots.
2021 — 2022 Lund · Stockholm, Sweden

R&D Researcher // Ericsson

  • Investigated object detection pipelines for AR/VR applications.
  • Explored deep-learning representations for semantic SLAM frameworks.
2020 — 2022 KTH · Stockholm, Sweden

M.Sc. Systems, Control & Robotics // KTH

  • Multidisciplinary Master’s focused on modern robotics — Applied Estimation, Computer Vision, Deep Learning.
  • Thesis with Ericsson: extending input modalities for deep feature extraction to improve robustness in Visual SLAM.
2020 Pi School · Rome, Italy

AI Fellow // School of Artificial Intelligence

  • Highly competitive, merit-based fellowship for top engineering professionals.
  • Engineered an AI-powered voice interface for call-centre triage using Natural Language Understanding.
2019 — 2020 Shanghai, China

Machine Learning Intern // Moseeker Inc.

  • Built an ML model to predict candidate employability from audio signals in large-scale interview datasets.
  • Designed data pipelines and audio feature-extraction workflows end-to-end.
2016 — 2019 Bologna · Shanghai

Double B.Sc. Automation Engineering // UniBo & Tongji

  • Rigorous dual-degree across Bologna and Shanghai — engineering with a global, cross-cultural lens.
  • Two theses: CNNs for proportional myocontrol (decoding muscle signals for prosthetic control); audio-signal classification for predicting candidate employability.
02 StackTools I reach for
C++
Python
ROS
Jetpack
SLAM
Machine Learning
OpenCV
Docker
Git