cv

Senior AI Engineer

PwC Italy - AI Center of Excellence

Building generative AI solutions and teaching AI programs for clients and partners.

• Led a small team delivering an open-source DiT-based video generation suite, optimizing inference with post-training quantization and step distillation.
• Co-designed and taught AI training modules for PwC Italy and university partners, including hands-on labs for executives and graduate students.
• Developed and improved retrieval-augmented generation pipelines, including efficiency tuning, retrieval and generation improvements, and an internal LLM evaluation pipeline.

Visiting Research Assistant (Master's Thesis)

Queen Mary University of London - Centre for Multimodal AI

Research on forensic techniques for synthetic image detection and origin attribution.

• Master's thesis "Forensic Techniques for Synthetic Image Detection and Attribution."
• Co-authored "Are CLIP Features all you need for Universal Synthetic Image Origin Attribution?" earning Best Workshop Paper at ECCV 2024.
• Ideated, implemented, and trained all experiments; conducted literature review and paper writing.

Research Assistant

University of Florence - Media Integration and Communication Center

Vision-language research on diffusion-based augmentation for cultural heritage datasets.

• Published "Diffusion Based Augmentation for Captioning and Retrieval in Cultural Heritage" (ICCVW 2023).
• Built a diffusion-based multimodal augmentation pipeline improving image captioning and multimodal retrieval.
• Developed and trained experiments and contributed to paper writing.

Research Assistant (Bachelor's Thesis)

University of Florence

Crowd counting research using domain-adapted CNNs on thermal imagery.

• Thesis "Convolutional Neural Networks for Object counting in thermal imagery."
• Leveraged domain-adapted YOLOv3 features to improve privacy-preserving crowd counting and cross-domain generalization.

Product Owner & Software Developer

Hermes Trade S.r.l.

Student worker leading ERP/CRM development across the software development lifecycle.

• Designed and led the development of an ERP/CRM application as backend developer and data engineer.
• Managed full software development lifecycle while working part-time during studies.

M.Sc. in Artificial Intelligence

University of Florence

Artificial Intelligence

• Machine Learning
• Deep Learning
• Computer Vision
• Natural Language Processing
• Reinforcement Learning

Bachelor in Computer Science Engineering

University of Florence

Computer Science Engineering

• Algorithms
• Data Structures
• Computer Architecture
• Operating Systems
• Databases

Best Workshop Paper Award, TWYN @ ECCV 2024

ECCV 2024

Awarded for "Are CLIP Features all you need for Universal Synthetic Image Origin Attribution?"

Are CLIP Features all you need for Universal Synthetic Image Origin Attribution?

ECCV Workshops

Best Workshop Paper Award at TWYN @ ECCV 2024; employs foundation model features for open-set synthetic image origin attribution.

Diffusion Based Augmentation for Captioning and Retrieval in Cultural Heritage

ICCV Workshops

Diffusion-based multimodal augmentation pipeline improving cultural heritage captioning and retrieval.

Deep Compression

• Reproduced the three-step compression pipeline on ImageNet subset and MNIST.

Deep Learning Applications

• Built end-to-end implementations of research papers for coursework and experimentation.