co-founder · UNaIVERSE ↗
peer-to-peer network for human-ai agents.
masaccio — developer & ai researcher
tommaso guidi
i'm tommaso, but everyone calls me masaccio. i study artificial intelligence and how to make it learn over time. i like big challenges and hunting for non-trivial solutions to complex problems; ideally, i also like finding them. a necessary but not sufficient condition for that to happen is looking at reality from multiple points of view, and being willing to change perspective.
raised in florence, based in siena (italy)
the long version
PhD candidate in Smart Computing (39th cycle, University of Florence, joint program with Pisa and Siena), doing research at SAILab, the Siena Artificial Intelligence Lab, under the supervision of Marco Gori, within the collectionless AI team. My work is about learning over time: recurrent models, state-space dynamics and collectionless AI, aiming at agents that learn from their environment without collecting data. In spring 2026 I was a visiting researcher at TU Darmstadt, in Kristian Kersting's AI & ML lab. Before the PhD: MSc in Artificial Intelligence and Automation Engineering in Siena (110/110 cum laude, thesis on instance-level recognition), BSc in Electronics and Telecommunications Engineering in Florence (thesis on hieroglyph segmentation with deep neural networks), research internships at IFAC-CNR on AI for medical imaging and cultural heritage, and two years as AI engineer at AIgea, developing the deep learning engine of DeepMammo for breast cancer detection in screening mammography. Co-founder of UNaIVERSE, a peer-to-peer network for human-AI agents.
working experience
soon
what's next
unstaged, open to interesting problems.
2026–
2025–
research fellow · DIISM, Università degli Studi di Siena
research fellowship on the theory of learning and reasoning without access to data (project CONSTR, FAIR extended partnership, PNRR), renewed in 2026.
2023
research collaborator · IFAC-CNR, Istituto di Fisica Applicata Nello Carrara
NAVIGATOR project (funded by Regione Toscana): computer aided detection of cancer in medical imaging.
2022–2024
ai engineer · AIgea S.r.l.
development of the DeepMammo model for breast cancer detection in screening mammography.
2021–2022
research intern · IFAC-CNR, Istituto di Fisica Applicata Nello Carrara
ai applications in medical imaging (cad of kidney, lung and prostate cancer) and in cultural heritage.
2019–2022
barista, delivery rider, waiter · various, Florence area
service jobs through the university years: Bar Collo (Cerbaia), Winelivery (Florence), Trattoria da Bule (Bargino).
research & education
learning over time: recurrent and state-space models, lifelong and collectionless ai; agents that learn by interacting with their environment without collecting data collectionless ai ↗
2023–
PhD, Smart Computing (XXXIX ciclo) ↗
joint program with Pisa and Siena; research at SAILab (University of Siena), supervisor Marco Gori.
↳ visiting researcher TU Darmstadt, AIML lab (Kristian Kersting) ↗, may–jul 2026
2021–2023
MSc, Artificial Intelligence and Automation Engineering
Università degli Studi di Siena · 110/110 cum laude
thesis: Pitfalls in Instance Level Recognition
2017–2021
BSc, Ingegneria Elettronica e delle Telecomunicazioni
Università degli Studi di Firenze · 108/110
thesis: hieroglyph segmentation with deep neural networks
2012–2017
Maturità scientifica
Liceo Scientifico Niccolò Rodolico, Firenze · 100/100
publications
cooking
phd thesis on Foundations and Directions for Lifelong Learning Agents
staged, not pushed yet.
2026
Perpetual Generation: Online Learning of Linear State-Space Models from a Single Stream
ICANN 2025 (Springer)
M. Casoni, T. Guidi, S. Melacci, A. Betti, M. Gori
2026
Book Your Room in the Turing Hotel! A Symmetric and Distributed Turing Test with Multiple AIs and Humans
arXiv:2603.18981
C. Di Maio, T. Guidi, L. Quarantiello, J. Bell, M. Gori, S. Melacci, V. Lomonaco
2025
Back to Recurrent Processing at the Crossroad of Transformers and State-Space Models
Nature Machine Intelligence 7(5)
M. Tiezzi, M. Casoni, A. Betti, T. Guidi, M. Gori, S. Melacci
2025
Generative System Dynamics in Recurrent Neural Networks
IJCNN 2025
M. Casoni, T. Guidi, A. Betti, S. Melacci, M. Gori
2025
A Unified Framework for Neural Computation and Learning Over Time
ECAI Workshop ML meets Differential Equations (PMLR)
S. Melacci, A. Betti, M. Casoni, T. Guidi, M. Tiezzi, M. Gori
2025
UNaIVERSE: A Peer-To-Peer Network For Human-AI Agents
preprint
S. Melacci, C. Di Maio, T. Guidi, M. Gori
2024
Pitfalls in Processing Infinite-Length Sequences with Popular Approaches for Sequential Data
ANNPR 2024 (Springer)
M. Casoni, T. Guidi, M. Tiezzi, A. Betti, M. Gori, S. Melacci
2024
Hyperspectral Imaging and Convolutional Neural Networks for Augmented Documentation of Ancient Egyptian Artefacts
Heritage Science 12(1)
C. Cucci, T. Guidi, M. Picollo, L. Stefani, L. Python, F. Argenti, A. Barucci
2024
On the Resurgence of Recurrent Models for Long Sequences: Survey and Research Opportunities in the Transformer Era
arXiv:2402.08132
M. Tiezzi, M. Casoni, A. Betti, T. Guidi, M. Gori, S. Melacci
2023
Egyptian Hieroglyphs Segmentation with Convolutional Neural Networks
Algorithms 16(2)
T. Guidi, L. Python, M. Forasassi, C. Cucci, M. Franci, F. Argenti, A. Barucci
2022
Ancient Egyptian Hieroglyphs Segmentation and Classification with Convolutional Neural Networks
Florence Heri-Tech 2022 (Springer CCIS)
A. Barucci, C. Canfailla, C. Cucci, M. Forasassi, M. Franci, G. Guarducci, T. Guidi, M. Loschiavo, M. Picollo, R. Pini, L. Python, S. Valentini, F. Argenti
side quests
someday
next side quest
unstaged ideas, waiting for a free weekend.
2026–
masaccio.cv ↗
this website: one data source, site and cv both compiled from it. available as a reusable template.
2026–
lesgoski ↗
flight deal scanner: looks for cheap ryanair round trips, and sends a notification when it finds one. fastapi dashboard, self-hosted on a bare hetzner vps.