CV

Senior Data Scientist with 7+ years of experience building end-to-end ML systems across product, research and clinical settings. Based in Milan, Italy.

Skills

Machine learning
Deep learning, computer vision, time series modelling, LLMs, prompt engineering, explainability, edge AI, on-device inference
Multimodal sensing
Sensor fusion (IMU, EMG, RGB, audio), wearable AI, biosignal analysis
Programming
Python (PyTorch, TensorFlow, scikit-learn, Pandas, NumPy, OpenCV), C++, C#, MATLAB
Frameworks
Hugging Face, ONNX, TensorRT, ROS, Unity, TensorBoard
Infrastructure
Docker, Google Cloud Platform, Azure, Databricks, Git, Jira, VS Code

Experience

Senior Data Scientist, R&D

Schindler Group, Milan, Italy

  • Technical leader for computer vision and LLM projects in the R&D department.

Senior Applied Data Scientist

EssilorLuxottica, Milan, Italy

  • Technical lead in a joint initiative with Meta, delivering human activity recognition and scene understanding models on proprietary smart eyewear hardware.
  • Led collaborations across four academic research groups, owning technical direction and delivery across eight concurrent projects within a national grant program.
  • Owned design, development and deployment of on-device computer vision for consumer smart eyewear, including ego-action recognition, context awareness, human keypoint estimation and SLAM.
  • Led model optimisation and compression for embedded deployment, with up to 75% reduction in model size while meeting latency and accuracy targets.
  • Built a computer vision pipeline for eyewear component segmentation and colour and style matching against a product catalogue, combining U-Net segmentation and histogram-based colour matching.
  • Designed and evaluated prompt engineering pipelines using QWEN and Gemini for OCR-based prescription extraction that fills a clinical web interface from raw prescription images.
  • Turned product and business requirements into production ML pipelines that respect privacy, hardware, latency and EU AI Act constraints.
  • Built reproducible training, evaluation and deployment pipelines with PyTorch, Docker, GCP, Azure and Databricks.
  • Supervised two researchers. Contributed 7 patent filings and over 10 peer-reviewed publications.

Machine Learning Research Associate

Imperial College London, United Kingdom

  • Developed longitudinal time-series models (LSTM, autoencoders) for passive health monitoring and early change detection in people with dementia, using clinical data from NHS-funded trials.
  • Designed explainability and robustness frameworks with SHAP and stability analysis to support clinical interpretability and model validation.
  • Improved preprocessing pipelines for wearable sensor data, reducing missing data by 5%.
  • Contributed to a successful Wellcome Trust grant through clinical collaboration and research outputs.
  • Supervised 3 PhD students and 4 master's students.

Machine Learning Researcher

HES-SO, Switzerland

  • Built an augmented reality rehabilitation system with Unity and Microsoft HoloLens, using EMG and IMU signals for real-time movement tracking.
  • Developed sEMG gesture recognition pipelines focused on robustness and cross-session repeatability.
  • Segmented lung and breast cancer CT datasets at scale.

Education

PhD in Electronics Engineering and Computer Science

Queen Mary University of London, United Kingdom

  • Built multimodal robotic perception systems combining tactile, proximity and visual sensing within the National Centre for Nuclear Robotics programme.
  • Implemented ML and computer vision pipelines in Python, OpenCV, C++ and ROS. Internship at the University of Tokyo on object detection (Faster R-CNN, YOLO) for infrastructure inspection.
  • Mentored over 10 MSc students and represented a PhD cohort of 20 at departmental level.

MSc in Artificial Intelligence and Robotics

University of Rome La Sapienza, Italy. Final grade 110/110

  • Researched grasp recognition from sEMG signals for robotic hand control. Internship at HES-SO on sEMG and eye-tracking dataset acquisition and classification.
  • Published at an A* conference. Recipient of the Best Master's Thesis Award.

Awards

  • Data Visualisation Competition Winner, UK Dementia Research Institute
  • First Prize, CORSMAL Challenge (ICPR)
  • Japan Student Services Organisation Scholarship
  • Best Master Thesis Award
  • Programmer Award, VVV18 Winter School

Publications and patents

IEEE Access (2025), Nature portfolio journals (2024, 2023), ICRA (2020) and ICORR (2017), plus 7 patents filed in applied computer vision and wearable sensing. See the full list.