The work on our Zoeked product started 5 years ago as we were working to create a platform that would enable analytics on CAD systems similar to the types of analytics we can get for web pages and site using google analytics.
The sheer amount of picks and clicks performed by a CAD user on any given day far outweighed the typical web session for any website. Further more, the variation of functions used to engineer and design complex products was astronomical to say the least. Building a simple data centric solution continued to require massive amounts of server compute time and began to slow with the more users added.
Starting with Node.JS as the main development language for all the server side work, we rewrote the analytics and incorporated some machine learning frameworks to make better sense of all the data we were receiving. The concept is to teach Zoeked what a proper user CAD session looks like for given tasks on different setups and the algorithm will return the required information.
The result is a robust system that can run predictive analytics and provide insights into user training needs, hardware and software setups and build a list of useful sites, videos and learning content for each specific users based on their actual usage. The best part is that Zoeked gets better over time. the more data, the more helpful it becomes.
The Open Source Idea
The final thoughts we had on launching Zoeked as a product came from the extraordinary success of Redhat Linux and that was to launch Zoeked as an open source framework with select customers having access and being able to contribute to the framework. This will ensure that we can engage open source frameworks for reporting, security and overall build quality while still continuing to quickly evolve the solution with a massive community of developers.