Publishing Data-science Models.

WIDE IO is a company specialised in technologies for the dissemination of data-science models. Since December 2012, it operates a consulting organisation and help businesses to solve data and algorithmic problems that they face. We also develop as part of our R&D an open-source platform that helps scientists to publicise, commercialise scientific results.

Finding optimal solutions.

Acquiring algorithms is complex. Most software vendors explain that they all have best data-science solutions. But can they really all be the best ? Can they say so without knowing your data nor your competition ? In small specialised technical market, the right solution among the different available alternatives, needs to be evaluated precisely taking in account the specificities of your needs including platform, data pipelines and metrics. This can take time and it is not always possible to assess all solutions - but best effort is still recommended.
Once a good solution is identified, bespoke development may be required to ensure that critical points of a specific market are adressed.
We are used to face this kind of problems - we are available as consultants to help you evaluate different models and integrate them into your next product.

Our job is to help you to build the right technology.

Our methodology

  • Passionated team

    Our team share a culture made of discrete mathematics, algorithmic, data science, software engineering. We generally believe that the right intelligence at the right place can help to make the world a better place by identifying optimal solutions to well-defined problems. Computer simulations can help to make the right assumptions and help you to understand the limit of their applications. Our team comes from academia and used to speak conferences and meetups. We like to exchange with other specialists, and we are used to work in team.

  • Flexible offers

    We offer to senior consultants priced by the day. It is particularly convenient for project with short time line - or that have already been planified. When the project to be executed incorporate execution risks, after an initial feasibility assessment, we are also happy offer to our customers the possibility to define a result-driven projects: we offer to share the risks of the project by engaging on numerical targets.

Our customers

Digital Catapult

We are working with Digital Catapult on the establishment on data-model that can scale at the size of the planet and that can serve as the key to future internet-scale interoperability issues. We have been working with them on different technologies involving notably graph engine and the semantic web .

Queen Mary University

In 2015, our team has been working with Queen Mary University to produce commercially viable software in the realm of video-analytics. While the product description is still confidential, it can be said that our team has entirely rewritten the program into modular and fast C++-11 code. Researchers have now high-quality code that they can integrate in the commercial software.

Semblent Ltd

Until 2014, WIDE IO Consulting has worked with Semblent Ltd and has build for them the analytic software that has allowed tSemblent to find a first round funding giving Semblent. During the year of contracting 2015, WIDE IO LTD has helped Semblent to explore text-analytics solutions and to design graph-mining solutions.

Techmarketing & LIDL

In 2014, can make a "Kiss Detector" for Valentine day ? Yes we can ! We have helped the Israel-based company Techmarketing with the comparison, integraton and adaptation of action-recognition algorithms for a project related to an advertising campaign ran for LIDL all over the globe during Febuary 2014. Our research has started with a bibliography on this topic, and with a collaboration with members of the computer vision team of Imperial College. Consequently, we have implemented a state-of-the-art action recogniser, evaluated on hand-made datasets, and optimised it both in terms of accuracy and performance by taking in account some application-domain specific knowledge. The result was a succesful adverstising campaign ran over 25 countries and used by 4M+ users.

Follow us

Contact us

io . wide @ hello