About
I am currently working as a Software Engineer at u-blox in Zürich, Switzerland and have just published my first Computer Vision paper at IEEE ICIP. I graduated with a double degree in Computer Science and Engineering at Polytechnic University of Milan and Wrocław University of Science and Technology. I am always seeking new challenges and opportunities to grow.
Work Experience
u-blox
- Leading the design and architecture of a global GNSS monitoring project, overseeing real-time data collection, network transmission, and cloud-based logging and visualization.
- Developing, enhancing, and maintaining C++ drivers for market-leading GNSS receivers on Android Automotive and Windows, improving thread-safety, and streamlining testing setup and management by introducing Ansible.
- Active member of the AI Center of Competence, providing expertise and support for AI projects across the company.
Imagination Technologies
- While working in a two-person research project, designed and implemented a mobile GPU simulator based on the gem5 project.
- Designed and maintained high performance GPU simulation models written in SystemC for the PowerVR graphics processors.
Nokia
- Worked in a R&D unit on the implementation and maintenance of the baseband service for new 5G solutions using C++ and Python.
- Established myself as the go-to person in the domain testing area. The improvements I proposed to the pytest-based testing framework brought measurable performance benefits and were recognized at the functional area level.
Intel
- Worked on the development of the driver for the cutting edge Infrastructure Processing Unit, IPU ASIC E2000 (code-named Mount Evans), which is a collaborative achievement of Intel and Google Cloud.
- Received managerial recognition for the contribution to the development of unit, component, and integration tests using the Google Test framework.
- Actively verified software using the Simics virtual platform.
Education
Polytechnic University of Milan
Faculty of Electronics, Information and Bioengineering, Double Degree Program with Wrocław University of Science and Technology.
Thesis: "SE3D: A Framework for Saliency Method Evaluation in 3D Medical Imaging".
Wrocław University of Science and Technology
Specialization: Applied Computer Engineering in Medicine, Faculty of Information and Communication Technology. Graduated with Excellent (A+).
Thesis: "Deep Learning Model for Object Detection in Medical Imaging".
Wrocław University of Science and Technology
Specialization: Applied Computer Engineering in Medicine, Faculty of Electronics.
Thesis: "Story Maker: Mobile Application for Automatic Generation of Video Stories Based on Files Specified by the User".
Autonomous University of Madrid
Computer Science Engineering, Faculty of UAM Polytechnic School.
Skills
Check out my latest work
I've worked on a variety of projects, including computer vision, natural language processing, algorithmic optimization, and mobile development. Here are some of my favorites.
Weakly Supervised Alzheimer's Disease Detection from 3D MRI Scans
A proof-of-concept weakly-supervised 3D CNN model for Alzheimer's disease detection from 3D MRI scans. The model uses 3D Grad-CAM to generate masks corresponding to regions affected by the disease.
Travelling Salesman Problem
A C++ program solving the Travelling Salesman Problem using exact, local search, and population-based algorithms.
Story Maker
A mobile application written in Flutter for both iOS and Android platforms. The app automatically generates short, story-like videos by selecting the best moments from user-specified files.
I love research
Till date, I have published 1 paper in a top-tier conference. It is a great way to share my research with the community and get feedback from experts. I hope to continue this trend in the future.
- S
IEEE ICIP
SE3D: A Framework For Saliency Method Evaluation In 3D Imaging
Abu Dhabi, United Arab Emirates
Paper introduces new metrics to evaluate saliency methods for 3D CNNs and modifies ShapeNet, ScanNet, and BraTS datasets for benchmarking. The analysis reveals current 3D saliency methods fall short in performance, highlighting the need for advancements in the field.