We love AI and...

  • Help you collect and build equity in your data
  • Train a huge AI on all the data to learn your business
  • Use this AI to solve all your problems

Yes, one AI for all your data and problems.

Apollo statue


Check out our campaign with OVH!

View post on Twitter


Unbox AI process

Principal investigators

Prof. Gunnar Carlsson photo

Prof. Gunnar Carlsson

Carlsson is Professor of Mathematics Emeritus at Stanford University, and with 42 years of research experience, Carlsson is one of the most renowned mathematicians in the world and one of the founders of Topological Data Analysis. He is also president and co-founder of the artificial intelligence platform Ayasdi.

Rickard Brüel Gabrielsson photo

Rickard Brüel Gabrielsson

Brüel-Gabrielsson is a Stanford graduate and MIT researcher. He is an acclaimed researcher and a serial entrepreneur with over 10 years of experience of using tech and AI to transform industries.

We’re the world’s leading AI co-investor

Venn diagram: AI + private equity = creating value
Altor logo

ALTOR are world experts at Private Equity.
They manage EUR 8.3 billion with an excellent track record.


We’re world experts at self-supervised learning for businesses. We develop state-of-the-art natural-language-modeling inspired AI models for consumption and behavior.

This includes a lot of engineering and coding. We work with some of the biggest organizations in the world and our solutions have been used at scale by millions of people.

Keywords: Self-supervised and unsupervised learning, NLP, NLM, vision, contrastive learning, Transformers, GPT3, BERT, T5, Electra, CLIP, BART.


R. Brüel Gabrielsson, Chris Scarvelis
arXiv preprint (2022)
R. Brüel Gabrielsson
NeurIPS (2020)
R. Brüel Gabrielsson, B. Nelson, A. Dwaraknath, P. Skraba, L. Guibas, G. Carlsson
AISTATS (2019)
R. Brüel Gabrielsson and G. Carlsson
ICMLA (2018)
R. Brüel Gabrielsson, V. Ganapathi-Subramanian, P. Skraba, L. J. Guibas
Computer Graphics Forum (2018)
G. Carlsson and R. Brüel Gabrielsson
Abel Conference (2018)

One AI for all your problems

Our AI unlocks the synergies in your business, breaking the one model per function paradigm.

  • Pool all data
  • Train a single model
  • Plug in all company functions and enjoy the synergy effects
Unbox AI paradigm

Self-supervised learning

Our AI learns through observation rather than from a teacher. The meaning it finds is relational rather than abstract. Together, these lead to superior performance.

Example: learning what is a dog using labeled data versus through observation


We've had our hands in most AI-related problems and find ourselves coming back to this insight: the power of AI allows for more general problem formulations, pooling of data, and solving more tasks using less customization. We avoid wasting time on problems that are soon solved by a better, more general problem formulation; instead, we go straight to the core of AI.

We're commercial and in addition to research we also implement and deploy our model. Our team includes brilliant engineers and programmers, covering ML, data, dev-ops, and web.