Experience

MyAv

• Italy
• IT Services and IT Consulting
• 1 - 100 Employee

• Computer Vision and Deep Learning Engineer

  • Feb 2021 - Present

  Build state-of-the-art deep learning and computer vision algorithms on resource constrained edge devices. Train deep learning models to perform visual recognition tasks. Optimize deep neural networks and the associated pre/post-processing to run efficiently on edge devices. Continuously improve deep neural network performance on edge devices, including accuracy, speed and power consumption. Research the latest trends and papers, implement and experiment with new ideas to improve performance. Collaborate and work with other teams such as data mining and labeling teams.



Università degli Studi di Salerno

• Italy
• Higher Education
• 700 & Above Employee

• University Research Assistant

  • Jun 2020 - Dec 2020

  Development of an intelligent system for human emotion classification. Training of two audio classifiers: A Deep Neural Network based on extracted audio feature and a Convolutional Neural Network (YAMNet) based on Mel-Spectrograms. Comparison of two different fusion strategies, late fusion and early fusion. Each audio classifier was fused with an existing Recurrent Neural Network classifier based on video frames. Development of an intelligent system for human emotion classification. Training of two audio classifiers: A Deep Neural Network based on extracted audio feature and a Convolutional Neural Network (YAMNet) based on Mel-Spectrograms. Comparison of two different fusion strategies, late fusion and early fusion. Each audio classifier was fused with an existing Recurrent Neural Network classifier based on video frames.



A.I. Tech srl

• Italy
• IT Services and IT Consulting
• 1 - 100 Employee

• University Internship

  • May 2017 - Jul 2017

  During the University Internship in A.I. Tech I worked on my Bachelor's thesis: "Project and developing of a web application based on Django". I developed a dashboard in Django, working on both front-end and back-end. The dashboard receives information in JSON format from several applications. Each application was executed on a different device - usually it was a security camera. The dashboard has to process and save the JSONs in a database. Then, when users request them, the dashboard has to aggregate those information and it has to show relevant charts. The data were about the people flow in an area of interest.