• Built code pipeline deployment (Green/Blue) for new microservices by applying terraform blueprint and finished ahead of feature’s release schedule• Reduced CPU utilization by 20% and resolved heap exhaustion for core business file transfer systems, through major version upgrade, data migration and heap tuning• Using Python, created system patching automation, eliminating hours of recurring manual work and allowing the team to focus on more impactful tasks• Created and modified Terraform modules for database engine upgrade , improving resiliency for 40+ RDS clusters• Planned and executed Data Migration that involved over 35 millions datasets for one of the main business services by using Mysql queries and AWS Aurora• Assisted and mentored other team members on Terraform modules, Ansible playbooks, and Python automation scripts on a daily basis• Earned awards for having the most growth within the team and most willing to solve complex challenges

• Using AWS EventBridge, Lambda functions and S3 bucket, automated monthly payroll reporting, reducing hours of toil time• Refactored and upgraded the code bases on over 100 AWS State machines and Lambda functions to avoid deprecated runtime in under a month• Using Terraform and AWS SSO, set up permission group and its associated IAM resources for teams• Built Identity mapping function in Python and AWS Lambda which maintains the system consistency between Okta access permission groups and drupal admin groups• Built internal API Microservices and maintained API documentation using swagger and JSON:API standard that allow hundreds of internal users to access admin dashboard through Okta

Develop web and mobile app for USAhello:• Created new models and tables to store hundreds of new immigrant user’s personal information through PostgreSQL and Django• Reconstruct search engines and enable new filter functionalities by using Haystack, Solr and Elasticsearch

YWCA – Natural Language Processing for incident reports • Using Bash script, MongoDB and AWS EC2, set up cron jobs for database backup and automated caching system, which improved the overall resiliency and performance of the system.• Enabled rating summation and auto-complete functionalities, enhancing user experience and allowing the AI powered system to assess the risk level of incidents for YWCA• Built and fine-tuned multi-class classification models to detect high risky incidents at over 75% accuracy with under 1000 and imbalanced datasets (Scikit-learn, Tensorflow and Pytorch)