Skip to content

Eugenio Giuseppe Nerelli

PhD Student in Information and Communication Technology

Born 1999 · Italian (Native) · English (C1) · Spanish (C1)

Professional profile

  • PhD researcher developing a multi-agent system for automated live presentation control, coordinating the architecture of an MLLM-based orchestrator with real-time video processing and pose estimation running on consumer hardware (M4 base, 6GB VRAM GPUs).
  • Deep learning practitioner with hands-on PyTorch experience: reproduced SOTA manuscript segmentation with Mask R-CNN, reaching over 88% Pixel IoU on the DIVA-HisDB dataset across 50K training patches.
  • Built an automated benchmarking framework comparing optimization solvers on graph-based resource allocation problems, with models in constraint programming (MiniZinc) and Answer Set Programming (ASP).
  • VR and UX researcher: designed and ran controlled experiments on interaction methods and UI paradigms in immersive environments, using standardized QoE metrics and statistical analysis with more than 50 participants.
  • Two peer-reviewed publications, with a third paper under review.

Experience

Researcher · University of València

Apr 2025 – Sept 2025
  • AI Director project: coordinated a multi-agent architecture for automated presentation control, using a multimodal LLM orchestrator with real-time pose estimation and camera direction.
  • Extended Vision-Language Model capabilities to process pre-recorded video streams alongside real-time webcam input.

UX Research Intern · University of València

Oct 2023 – Jun 2024
  • Designed and ran a controlled VR study comparing interaction methods (ray-casting, teleportation, hand tracking) and UI paradigms (world-space, screen-space) for cultural heritage exploration.
  • Developed the experimental protocol with standardized questionnaires (SUS, UEQ, Presence) and quantitative metrics, recruiting and testing more than 50 participants.
  • Modernized legacy Unity/XR code and wrote Editor tooling for automated scene configuration and data logging.

Education

PhD in Information and Communication Technology

Sept 2025 – Present

University of València

  • Research line: integration of generative AI models (LLMs and VLMs) in Virtual Reality environments, to automate complex interactions and improve the experience.

Master's in Artificial Intelligence and Cybersecurity (Double Degree)

2021 – 2024

University of Udine (Italy) and University of Klagenfurt (Austria) · 107 / 110

  • Core focus: deep learning, computer vision, natural language processing, automated reasoning (constraint and logic programming), AI ethics, data science, cybersecurity fundamentals.
  • Studies abroad: a semester in Klagenfurt (Austria) and an Erasmus in València (Spain) focused on VR and UX.

Bachelor's in Internet of Things, Big Data and Web

2018 – 2021

University of Udine · 94 / 110

  • Focus on IoT, big data and web development.

Publications

  1. Immersive Reconstruction of Cultural Heritage: a Framework for investigating QoE aspects of interaction in Virtual Reality

    E. G. Nerelli, Á. Egea Benavente, J. Segura-Garcia & M. Montagud

    2025 17th International Conference on Quality of Multimedia Experience (QoMEX), Madrid, Spain

    2025 · doi:10.1109/QoMEX65720.2025.11219959

  2. Reconstrucción virtual inmersiva de teatros del Siglo de Oro desaparecidos: La Casa de Comedias de l'Olivera (València)

    E. G. Nerelli, Á. Egea Benavente, J. Segura-Garcia, J. A. De Rus & M. Montagud

    VII Congreso Internacional de Humanidades Digitales Hispánicas (HDH), València, Spain, 22-24 Oct. 2025. ISBN 978-84-9133-830-7

    2025 · doi:10.7203/PUV-OA-9788491338307

  3. Research paper on VR UIs and interaction methods

    E. G. Nerelli et al.

    Full paper comparing UIs and interaction methods in VR, in revision before submission.

    In preparation

Technical skills

AI and deep learning
PyTorch, TorchVision, CUDA, Scikit-learn, Hugging Face Transformers, NumPy, Pandas
AI specializations
Computer vision (CNNs, Mask R-CNN, object detection), Vision-Language Models, Multimodal LLMs, On-device model deployment, Prompt engineering
Automated reasoning
Constraint programming (MiniZinc), Answer Set Programming (ASP)
Programming languages
Python, C#, R, Java, HTML/CSS
Platforms and tools
Unity Engine, Git, Three.js
Specialized skills
XR development (OpenXR, Meta SDK), Unity Editor scripting
Research methodologies
Experimental design, Hypothesis testing, Quantitative and qualitative data analysis, Statistical analysis (SPSS, R), Scientific writing, UX evaluation (SUS, UEQ, Presence)

The open source work is documented under projects. To get in touch, the addresses are on the contact page.