Develop and maintain MLOps for machine learning teams in the company

- Implemented a stream pipeline to ingest geolocation data into Elasticsearch with a geo coordinated map on Kibana to visualize the data in near real-time (Java, Flink, Elasticsearch, Kibana). - Automated the establishment of Data Lake platform with autoscaling functionality and developed several data pipelines to aggregate and cleanse the unstructured distributed data in it (Hadoo family, Cloudbreak, Kylo/NiFi). - Developed both Batch and Streaming pipeline to process large data for ingesting and modelin purpose (Java, Storm, Flink, Spark, Hive)

- Developed the new features and fixed the bugs of the system automation software called IPcenter (Java/Spring, Hibernate/MySQL, ActiveMQ). - Reduced the manual tasks by making it automatic to run Unit Test and Integration Test after each code push (Java, JUnit, Selenium, FindBugs, Jenkins). -Improved the code quality by making a rule that the code must be merged after the reviewers’ approval and after passing the tests (Git, Bitbucket).

- Had resolved various difficult technical issues that other support engineers escalated. Also, filed the found bugs in a manner to help the developers to find the cause and resolution (Java, MySQL/PostgreSQL, LDAP, AWS, etc). - Had fixed some double-byte character related bugs (Java, Javascript). - Integrated Confluence with user dictionary to improve search relevancy (Java, Lucene).

- Customized the existing REST API server to update a certain customer’s data in real time to meet their requirement to close the deal (Java/Spring, Cassandra, ActiveMQ). - Established a distributed Linux server infrastructure that meets high performance requirement - 500QPS + 30ms response time (Nginx, Tomcat, Cassandra, ActiveMQ). - Automated the server management tasks, which dramatically reduced manual operation workload using Chef, which enabled to share among other members in North America the up-to-date information of how to establish the infrastructure (VPS, Linux, Chef, Ruby)