Implemented Edge Connection as a service to connect between any colocation and customers’ site, integrated automatic connection with REST API, improved connection latency

Built logging systems to track all the CRUD operations, and then saved logs on both MongoDB and Elasticsearch via Logstash plug-ins installed in Docker, provided assistants for the system debugging

Implemented distributed system like for Kubernetes cluster and KubeEdge, supported metadata synchronization between cloud and edge

Developed frontend UI using JavaScript and React with react-jsonschema-form, implemented UI components including modal creation and form validation

Upgraded Ubuntu servers and set up Yarn services, improved site reliability

Developed Airflow and applied it for scheduled jobs, provided a more organized framework for running factdata analysis

Managed CloudWatch, API Gateway running on Amazon Virtual Private Cloud and tracked logs for live issues

Used Git version control tool to coordinate project status and procedures and manage different Git versions on AWS CodeCommit

Designed A/B test frames to test outcomes under different incentives for similar testing groups, cleaned up more than 3000 contacts using data engineering methods

Utilized Machine Learning models to predict index related to paid-search digital market of different industries, including Logistic Regression and Linear Regression; the model predicted a positive prospect for entering the new market

Constructed regression and cluster models to evaluate whether the industry is suitable for applying new Blockchain technology