About
Who we are
We are a carefully chosen collective that have worked with each other for years (in some cases, decades) on researching and developing technology for machine learning and AI, working with a wealth of data sources and problems.
We have also taught extensively, building teaching and practical training materials that are in use by top universities in the UK, and elsewhere.
Team CVs and academic profiles are available on request.
What we do
We have broad experience in computational approaches to problem solving. Our members have experience in modern machine learning, pattern recognition and signal processing.
Our wide network of collaborators includes technology innovators, and entrepreneurs who have built and scaled both teams and deep technologies from the ground up.
We can help you do the same. Do you need to keep pace with the competition in your industry? Do you want to get an AI team up and running? We can help.
Why "snug"?
In addition to trustworthiness, the term "snug" refers to a form of linear representations for vectors in finite dimensions. "Snug" representations provide greater freedom in designing certain algorithms for processing signal and visual data than standard (orthogonal) basis representations.
Practically, snug representations can also reduce computational costs of training and the quantity of data required to achieve the same level of performance as end-to-end training.
