Eric George

I created a builder image and associated OpenShift templates for deploying Dropwizard REST services into an OpenShift cluster.

This allows for the creation of CI/CD pipelines for this type of application in a Kubernetes context.

As part of my work building an enterprise application for a healthcare software company, I set up the security for the app to utilize the existing LDAP server for authentication.

I designed the LDAP schema to meet the needs of the new project, and also be compatible with the structure of the existing LDAP server, which was relied upon by other legacy apps. I configured Spring Security in the new app (which was EJB-based) to use LDAP.

Once the code was complete, I created scripts and LDIF files to bootstrap new installations as well as automate Dev and QA deployment with a standard set of test and demo user credentials.

I designed and built an Android application that accepted card-swipe payments for taxis on a tablet mounted in the backseat. The application integrated with the taxi meter using raw serial communications to capture trip duration information.

During the trip, the tablet would play video advertisements. It would receive push updates of both the application as well as the advertising content.

We had several web applications, some running Spring Boot and one using EJBs and Spring Security. In order to make these applications function seamlessly together without users having to log into each, I implemented a single sign-on solution to consolidate the authentication.

I configured Keycloak as my identity provider, using an existing LDAP server for its data store. Since all of the applications used Spring Security, I was able to configure that using the appropriate OpenID adapter.

One of the challenges with this migration was that the EJB application needed to maintain the ability to authenticate directly against LDAP. This was to allow for a smooth migration in case the new authentication scheme did not work.

I worked for a startup that had to process a huge number of incoming emails. The emails would be sorted by one script, pushed to various servers where a cron job would pick them up and create database records for some of them. Then a similar cycle would send the data along further.
With all of the cron jobs polling and checking, it took over 30 minutes to process something that took only a few seconds of computer time.

I saw there was a better way, so I built an event management system with Java that could trigger a task remotely when it was needed, eliminating all of the wasteful polling and waiting. The product used Java-RMI to communicate between servers and XML to define the jobs (both new technologies at the time).

I eventually licensed the product to a company who made software to synchronize data between disparate relational databases. We enhanced the product to allow for embedded Jython scripts and wrote connectors for the transaction logs of major database systems.