Resume
Table of Contents
Employment
Jun 2023 – Present · Machine Learning and Computer Vision Scientist, EssilorLuxottica, Milan, Italy
- On-device computer vision for smart eyewear: context recognition, ego-action recognition (miniROAD), human pose and face keypoints (YOLO-Pose), and SLAM (ORB-SLAM3).
- Edge deployment and TinyML: 75% model size reduction with at least 80% accuracy.
- Lead collaboration with Politecnico di Milano and co-lead project with Meta on on-device, camera-based context recognition.
- Explore biomarkers from RGB eye imaging; initiate projects on non-invasive glucose monitoring and anaemia detection.
- Build reproducible pipelines in Python (PyTorch, OpenCV, Pandas) with Docker and GCP (Vertex AI, Buckets).
- Manage and supervise two researchers; drive IP and publications (7 patents submitted, 5 peer-reviewed papers; workshops at ICCV and IJCNN).
Jun 2021 – May 2023 · Machine Learning Research Associate, Imperial College London, UK
- Developed deep learning models (LSTM, autoencoders) in PyTorch and TensorFlow for personalised time-series and longitudinal clinical data (NHS) to detect health-related episodes in dementia.
- Improved data quality by reducing missing data by 5% within weeks.
- Proposed methods for explainability and robust learning from unreliable data using SHAP values.
- Contributed to securing a Wellcome Trust grant for the research group.
Feb 2020 – May 2020 · Computer Vision Intern, Tokyo University of Agriculture and Technology, Japan
- Implemented a TensorFlow multi-modal fusion model for object detection (Faster R-CNN, YOLO), fusing visual and tactile signals and resulting in two publications.
Jul 2017 – Jun 2018 · Machine Learning Researcher, HES-SO, Switzerland
- Applied machine learning and computer vision to healthcare:
- Augmented Reality training environment (Unity, Microsoft Hololens) for amputees, using MYO IMU data.
- Analysis of hand movement recognition repeatability with sEMG and ML classification.
- Image segmentation of >2000 CT lung and breast images for cancer pattern analysis.
Oct 2016 – Jan 2017 · Machine Learning Intern, HES-SO, Switzerland
- Worked on the MeganePro project: acquired, preprocessed, synchronised, filtered, and classified sEMG data using Random Forests.
- Published in a rank A conference; Runner-up Best Student Award.
Education
Ph.D. in Electronics Engineering and Computer Science
Queen Mary University of London, UK · Jul 2018 – Sep 2022
- Multi-modal robotic exploration for fracture detection in extreme environments (NCNR collaboration).
- Object detection (Faster R-CNN, YOLO) in TensorFlow for crack localisation from video.
- OpenCV pipeline for crack skeleton extraction and path planning via minimum spanning trees.
- Implemented in Python, C++, and ROS; used CNNs and Random Forests for tactile and proximity sensor data.
- Ph.D. representative for ~20 researchers; mentored three master students; selected for Google Get Ahead 2020; involved in STEM outreach.
M.Sc. in Artificial Intelligence and Robotics
University of Rome “La Sapienza”, Italy · Jan 2015 – May 2017 · Mark 110/110
- Machine learning for grasp recognition from sEMG for robotic hand control.
- Android application for vocal control of a Nao robot (client–server).
- WebGL videogame using Three.js and a project on Kuka manipulator pose control via mutual-information visual servoing.
B.Sc. in Computer and Automation Engineering
University of Rome “La Sapienza”, Italy · 2011 – 2014 · Mark 95/110
- Thesis on Izhikevich spiking neural network models.
Awards
- Best PhD Thesis, QUeen Mary University of London (2022) - Second prize.
- Data Visualisation Competition, UK Dementia Research Institute (2022) – First prize.
- CORSMAL Challenge, ICPR 2020 – First prize (team); computer vision and ML for container capacity estimation.
- Japan Student Services Organization Scholarship, TUAT (2020) – Scholarship for three-month internship.
- Best Master Thesis, University of Rome “La Sapienza” (2019).
- Programmer Award, VVV18 Winter School on Humanoid Robot Programming (2018).
- Swiss-European Mobility Programme and HES-SO Grant, HES-SO (2016).
Portfolio
For a complete list of projects, see the projects page.
Recent examples:
- CNNs for Alzheimer MRI classification.
- Online GAN-based video animation using first order motion models.
- Convolutional GAN in PyTorch for handwritten digit generation.
Technical Skills
Languages
- Python (PyTorch, TensorFlow, Pandas, Numpy, OpenCV, Seaborn)
- C++, C#, Matlab
Tools and Frameworks
- Google Cloud Platform (Vertex AI, Buckets)
- Docker, Git, Visual Studio Code
- Hugging Face, TensorBoard
- ROS, Unity
- TinyML for embedded model optimisation
Languages
- Italian (Native)
- English (Advanced)
- French (Intermediate)
- Japanese (Basic)
Publications
For the full list, please refer to my Google Scholar profile.
Selected publications:
- F. Palermo et al., “Advancements in Context Recognition for Edge Devices and Smart Eyewear: Sensors and Applications”, IEEE Access, 2025.
- F. Palermo et al., “TIHM: An open dataset for remote healthcare monitoring in dementia”, Scientific Data, 2023.
- F. Palermo et al., “Designing a Clinically Applicable Deep Recurrent Model to Identify Neuropsychiatric Symptoms”, NeurIPS R2C Workshop, 2021.
- F. Palermo et al., “Implementing Tactile and Proximity Sensing for Crack Detection”, ICRA, 2020.
- F. Palermo et al., “Repeatability of grasp recognition for robotic hand prosthesis control based on sEMG data”, ICORR, 2017.