Sustainability
Massive reduction of material waste — aggregate reused at full value instead of down-cycled or landfilled.
Material vision for the future
We read what recycled road material actually is.
Road construction devours mineral aggregate, and most of it is discarded or stockpiled because no one can cheaply tell what recycled material contains. Matterly AI is the software layer that classifies road-material composition in real time, so it is reused at full value instead of down-cycled or landfilled.
The mission
Safety
Detection of harmful material fraudulently used as building material, before it enters a road.
Research
State-of-the-art material vision: hyperspectral sensing fused with concept-aware vision transformers.
Chemistry meets machine vision
Hyperspectral vision beyond optical inspection, enhanced by spectroscopic filter solutions, combined with vision transformers that integrate automatically learned human concepts into the training process — research from RESOLV and the Research Center Trustworthy Data Science in Dortmund.
Two billion tons a year
Construction and demolition waste is produced at roughly 2,000,000,000 tons per year, and only a fraction is genuinely recycled. Regional construction, utility and imaging-hardware partners are already lined up around the first deployments.
People
Cofounders
Matterly AI is built by a scientist-founder and a professor with decades of experience in trustworthy AI and molecular spectroscopy.
Dr. Carina Newen
CEO & Founder
Carina is a Postdoc on Interpretable and Reliable Machine Learning at the Research Center Trustworthy Data Science and Security, TU Dortmund. Her research brings explainable AI, robust AI and uncertainty quantification together — the trust foundations behind Matterly's material-vision systems.
Prof. Dr. Martina Havenith-Newen
Cofounder
Martina is Chair of Physical Chemistry II at Ruhr University Bochum and a leading researcher in terahertz and high-resolution infrared spectroscopy. Her group at RESOLV advances the spectroscopic sensing science that helps Matterly read what recycled road material actually is.
Updates
Progress notes
Short posts on how the prototype, the data and the partnerships are developing.
The interest list is open
We opened the Matterly AI interest list today. If you work in road construction, recycling operations, imaging hardware or applied material research, leave your email and we will contact you when the prototype concludes.
Next updates will cover our hyperspectral data collection, first classification results on recycled aggregate, and the pilot partners joining us.
Interest list
Be first in line when the prototype is ready
Leave your email and we will contact you when the Matterly AI prototype concludes, with pilot slots, results and next steps.