Working with international data science teams can be a challenge – particularly without any unified infrastructure.  

We provided Faculty’s data science infrastructure to a global data company, which led to millions in savings, stronger collaboration, increased team productivity and higher employee retention and satisfaction.


A global data and information company with over 40 offices around the world.


The company had identified a lack of a unified data science infrastructure (or in some cases any infrastructure). Ownership was often devolved to separate IT teams, or inflexible and unpopular third party tools were used. This made it impossible to collaborate across regions, and difficult even within teams. The fixed nature of one team’s cluster meant that the team were paying for computation power even when it was not being used.

Faculty worked with the company to evaluate Faculty Platform across three international regions. The aim was to robustly test the technical capability of Faculty Platform using a series of data science projects, while proving that it can make teams more productive, able to collaborate more effectively and allow them to do things that were previously unachievable.


Faculty’s engineers installed Faculty Platform on the client’s Virtual Private Cloud on Amazon Web Services. The platform was integrated with the client’s authentication systems (a common third party tool called Okta) and client databases were integrated to allow secure access to existing data. Access was restricted to a whitelist of IP addresses.

The platform reduced dependence on data engineers throughout the data science workflow, by providing data science teams with ownership over their infrastructure and removing time-consuming boilerplate code. As a result, data scientists were able to create accounts and set up working environments within minutes. A consistent platform also allowed code and models to be easily adapted for use in multiple regions, and reproducible environments reduced overheads for effective collaboration.

Over the course of the engagement teams within the company undertook seven machine learning projects using Faculty Platform, ranging from an experimental project classifying audio calls to a predictive model for loan default rates. Teams noted a number of timesaving benefits from using Faculty Platform. For example, data scientists were able to launch hyperparameter searches, consisting of thousands of individual runs, in order to find the best possible model for a given project. Without the ‘Jobs’ feature on Faculty Platform, this would have taken days to complete.

Faculty Platform facilitated bespoke training events with participants often joining remotely from many regions. Teachers were able to share code and exercises in real-time and participants had scalable access to servers with preinstalled libraries and instantly reproducible environments. Internal upskilling events such as this led to improved team productivity, new categories of technology, and employee retention and satisfaction.


By adopting Faculty Platform the company will save over £3.8 million a year through reduced engineering resource spend.

The flexible and on-demand compute available with Faculty Platform is forecast to reduce spend by approximately £300,000 a year.

Data scientist time saved by the platform will deliver £7,200,000 a year in productivity savings.

Teams collaborated more effectively, both locally and across different countries.