Imporment projects / design - IaC decouple: * Break down IaC project to from mono-service to cohesion service * Reduce IaC change task reaction time: 30 mins per plan / apply to <5 mins * Migrate IaC tool form Terragrunt to Pulumi - Log flow refactor(in design): * Add Logstash as a middleware to filter log and handle alert which trigger by log * Separate Elasticsearch log index form different Team’s service and setting index expired day - Zero downtime observability: * Build up monitoring standards for key services with Developer, Customer Success and PM teams * Setting monitor and alert for key service Maintains cloud infrastructure and network routes between two cloud provider: - AWS, Alicloud - special network resource cross cloud: AWS direct connect, transit gateway, Alicloud VBR, CEN - Kubernetes platform Maintains observability components - ELK, fluent-bit - Grafana, Prometheus, alert manager - Jaeger(Tracing) - Cloud resource log: S3 log, etc. Handling production troubleshooting, alert and emergency issue Review Cloud’s bill and resource cost, develop cost down plans

AWS infra: - Make clear existing AWS infra architecture and convert to infrastructure as code(using cloudformation) - Unify the current infra architecture and clean out differences between projects - Build metrics monitor with grafana - Build Cronjob trigger and monitor with Rundeck Use Python Script and AWS boto3 SDK to interactive with AWS - Set up EKS and build Airflow into it, provide Datateam as a workflow platform - Use Helm to manage K8s resources and for deploy - Identify AWS alarms, infra troubleshooting and RD daily requests - Review the resource usage and talk about optimization options with RD - Compose the infra doc, analyze AWS bill, track uncommon increased fees with services DevOps: - Redesigned Drone CI scripts to fix building jobs conflict and clean out differences between projects - Solved the problem of Image caused by triggered building job at the same time - Watching final deployment status and link build error alert to Slack - Redesigned Dockerfile script to optimize Docker image size and build time - Reconfigure filebeat, logstash to carry more information to Elasticsearch - Optimize the log at cron job link to kibana, so that RD can quickly locate the error message of a single cron job - Migrate some project from EC2 to ECS, and convert ECS EC2 cluster to fargate Backend: - Import Laravel's Job Queue package into CodeIgniter - Assist the backend to import Monolog package to CodeIgniter and integrate with ELK

- Planned CI/CD pipelines and Kubernetes clusters - Dockerized and redeployed CMSs and registration event websites - Maintained legacy PHP(5.1~7.2) system, Linux servers, and MySQL servers

- Created gitlab-ci pipelines to deploy systems automatically - Refactored and cleaned legacy codes in API system through OOP to improve code reuse - Refactored the sending module for Line Message API to increase efficiency

- Maintained project(develop with Laravel), hotfixed bugs, developed RESTful APIs, and optimized data queue - Integrated data using JQ refactor to parse MongoDB data to TSV - Automated deployment procedures by coding gitlab-ci.yaml to process build and push Docker images to the registry