Stefano Masneri
Hello there!
I'm a Senior AI Engineer at Mediapro. I build AI that understands video and AI that generates it. For the full career story, see my CV.
After my master's I worked as a researcher at the University of Brescia. I was curious about what working abroad would be like, so in 2009 I took a six-month contract in Berlin. Those six months turned into six years, followed by three more in Frankfurt am Main. Nine years in Germany changed my perspective on a lot of things, and made me a different (hopefully better!) person.
In 2018 I moved once more, this time to Spain, and San Sebastián is home now. Since leaving Italy I've picked up four foreign languages, and a good number of programming ones too. I also obtained a PhD from the University of the Basque Country, where I researched collaborative, multi-user augmented reality experiences for education (dissertation, 5 MB · defence slides, 12 MB).
Outside of work, I spend my free time with friends and family, playing basketball, reading, travelling, and discovering Gipuzkoa by bicycle.
I also try to learn something new every day, and I'm especially fond of the web as a wonderful interactive medium for discovery and understanding. No blog, but I do maintain a now page with what I'm currently up to.
Work
Study
Projects
- AVATecH, 2009 – 2013
- Software for neuroscience, 2015 – 2017
- UFC fighter tracking, 2017
- Audience engagement, 2020
- MLOps on GCP, 2022 – 2023
- cleAR, 2023
- AR Geography, 2024
- Gaussian Splatting, 2025
- AI Production Studio, 2026 – present
- inEvent, 2011 – 2014
- TRACTION, 2020 – 2022
- RAG for banking, 2023
- BrandIA, 2025
Places
- Brescia, 2000 – 2009
- Berlin, 2009 – 2015
- Frankfurt, 2015 – 2018
- San Sebastián, 2018 – present
Papers
- 2008: 1 paper
- 2009: 1 paper
- 2010: 3 papers
- 2011: 5 papers
- 2012: 1 paper
- 2013: 1 paper
- 2014: 4 papers
- 2015: 1 paper
- 2019: 1 paper
- 2020: 6 papers
- 2021: 3 papers
- 2022: 5 papers
- 2023: 2 papers
- 2024: 3 papers
Swipe sideways to move along the timeline.
3D reconstruction using Gaussian Splatting and NeRF
Capturing real locations as photorealistic assets for LED-volume virtual production. Benchmarked nerfacto, splatfacto and MCMC against RealityCapture and Postshot, and shipped a desktop app that runs COLMAP pose estimation and reconstructs either locally or on Azure.
Brand recognition in football matches
Measures how much screen time each brand gets in a football broadcast. A YOLO detector finds the boards, a DenseNet201 sorts them into 52 brands, despite motion blur, a 40× class imbalance, and LED ads that change weekly. In production on Azure.
ARoundTheWorld: Collaborative AR for Education
ARoundTheWorld is a multiplatform collaborative AR geography application built on the cleAR architecture. It was evaluated with 44 students across three schools and represents the final paper of my PhD, validating that the architecture can produce applications that integrate seamlessly into existing school curricula.
A collaborative AR application for education: from architecture design to user evaluation
cleAR: an interoperable architecture for multi-user AR-based school curricula
Interactive, Collaborative and Multi-user Augmented Reality Applications in Primary and Secondary Education. A Systematic Review
For a closer look, head to the Travel page. And for the two places that are more than pins: the one I stopped travelling to and started living in — Donostia en datos, seven stories about a city that changes (in Spanish) — and the one I left to get there, Come cambia Brescia, ten stories about a province in numbers (in Italian).
Let's talk
Stalk my projects on GitHub, read my papers on Scholar and offer me a job on LinkedIn. And if you prefer the old-fashioned way, the address is right there — copy it and write me an email.