eServices Greece was established in 2009 in Athens, Greece. The original idea was to provide hosting services for small sized companies as a full service. We had identified that over 90% of small companies are unable to manage their own infrastructure or web presence since they dont employee any IT personel due to tight budgeting. Thus eSGR worked in filling the gap by providing a service spanning from the installation of the web app to basic server side management of the customers service.

If our goal was to provide a service that was relevant, reliant and secure but also affordable for a company that didn't have the budget for its own IT department we would have to rely in automating processes as much as possible and thus in early 2011 we started developing our in house solution for web application optimization server side. By mid 2012 we had our own optimization engine that would latter be named Halcyon.

Working on our tools proved to be extremely important very early in the development process. Before we would require over 2 weeks of setup and engineering work per client but close to the end of 2012 the optimization process had gone down to a couple of days for the most common CMS we were hosting. In early 2013 our optimization engine started doing a lot more work, it included some basic filtering for known security threats and a caching mechanism similar to memcached for micro caching.

The largest change in how we worked started happening in 2015 when we started getting bigger clients with a lot more requests and higher difficulty issues. We needed to re-think the whole process and we decided we should invest our time in developing ML systems to help with the management of our customer services. The first system was put in production in mid 2015 and it proved invaluable since it could parse our logs and identify issues a lot faster than any of our engineers were able to.

In mid 2016 we felt comfortable enough to allow our ML (by now called Halcyon internally) train on data collected from solutions provided by our engineering team with the end goal of allowing the system to provide assistance for new issues, By the end of the year we had already allowed the system to provide fixes for issues it would identify as viable at over 85%.

In early 2017 we finally made Halcyon public by allowing our customers to interact with it directly with their applications. In its then current state Halcyon was able to provide solutions directly for over 90% of the requests made to our support department without the need to fallback to an engineer. In mid 2017 to simplify the model we split off from Halcyon Soteria (WAF), Coeus (Fabric) and by the end of 2017 Mnemosyne(DB, KVS). During this time we also got our first client with over 100 employees which would be almost impossible to support without our Halcyon Engine.

In 2018 we managed to get Halcyon to get a site up and running fully optimized in less than 25 seconds, at this time it takes more time for our paying processors to accept a payment than to deliver customer services. We will continue extending and expanding our technology for years to come and we hope in the near future to be able to provide even better services in less time.