Integrated our CORE system with external systems such as iBon, Zendesk, and Bank System. Developed internal systems such as AML engine, task management workflow system. Improved testing performance and compiling time. Analyzed and optimized PostgreSQL statements. Decoupled important functions of the legacy CORE system to new micro services. Upgraded the legacy Erlang version to the community supported version.

Developed new functions for product cloud 2.0 by NodeJs. BGP Blackhole router automation. Rewrote script by Python and Integrated internal systems, reducing the time to implement from 90 mins to 2 mins.

Developed a new feature of the inventory system with Python, Django, and PostgreSQL. Migrated MongoDB replica set to a larger hard disk space without downtime. Migrated cloud service to re-designed infrastructure stack without downtime. Optimized distributed task queue system(Celery) for stability. Leverage distributed task queue system(celery) and AWS SQS service to solve Taiwan invoice state inconsistency issue due to data upload out-of-order. Re-designed infrastructure stack via AWS CloudFormation to improve stability and efficiency by 800% with 50% less of the budget. Reduced 80% of all reporting features' response time by normalizing un-structure JSON data in MongoDB to MySQL database. Built high availability and high-performance system with AWS architect. Built a data storage system via MongoDB replica set instead of S3 storage, reducing 91% data transfer time.

Analyzed the relationship between specific large seismic events and related data from the ionosphere, groundwater, and GPS data. Found the seismic source from lots of detectable stations around Taiwan. Developed an awesome high accuracy and excellent performance auto-picking program, reduced 87.25% working time. Overall, the detect accuracy had 60% for all seismic events, for the larger than ML3.0 seismic events the accuracy is up to 100%