• Architect and implement container-based test automation framework running on on-premises cluster using Docker Swarm to run parallel testing using Nginx as both the front end web server and load balancer facing the public and Python Django and Python Celery as backend server. • Build private Docker Registry, private DNS and private NFS and Samba server on Docker shared volumes on the on-premises cluster. • Design analytic tool using Apache Spark to analyze passing rate, failure patterns, etc. from test records accumulated up to 5 years old stored inside SQL database. The tool is running in cluster mode using Apache Mesos on a 5 nodes setup.

+ Architect and implement the test automation infrastructure model to yield individual components within a chassis-based products consisted of over 80 FRU components. Main features of the model: - Parallel testing of components within a chassis is supported - Multiple parallel communication sessions (e.g. SSH, Telnet and COM sessions) are supported and are logged. Operator can view the progressive log of the communication sessions during test. - All components together with the overall chassis have their own test status at the end of a test run and is generated with its own HTML test report. - Test status of components and chassis is sent to CM database in web-service real-time connection to facilitate yield analysis by CM. Real-time connection to Flextronics and Foxconn shop floor database is supported. - Test status and test report is also sent to a company owned centralized NoSQL database to meet the various needs, e.g. yield and throughput calculation, failure analysis. - Create web-based production real-time monitoring tool to monitor all the Unit-Under-Test (UUT) on production floor. Running status of UUT and its components can be viewed during test. + Design and implement SKU management system to manage SKU configuration for UUT customization through web service connection. - The tool can allow user to create SKU with different combination of hardware configurations, software versions and licenses. - Manufacturing can associate UUT serial number to a SKU via a web service call to the tool at configuration station where UUT is built and configured accordingly. - The tool is written using Python Django web development framework.