Refactored GitLab CI from linear stages to a DAG with aggressive parallelism, smart caching, and targeted artifacts. Moved I/O‑heavy performance tests to on‑prem storage near runners. Result: ~50% faster CI and ≥12 hours shaved from performance testing, significantly improving developer throughput and release reliability.
A streamlined pipeline that minimizes manual actions and enables effortless movement from local development to a fully managed cloud environment. It emphasizes reproducibility, traceability, and speed—so teams can start fast, iterate confidently, and ship models reliably.
An end‑to‑end Infrastructure‑as‑Code project that bakes a JMeter AMI, deploys an internal ALB fronting an EC2 Auto Scaling Group, drives synthetic load from a private JMeter runner, and ships a CloudWatch dashboard to validate scaling behavior in real time—all without Internet/NAT Gateways.
A complete, production‑style path from source to running services on AWS. Microservices (frontend, backend, and data crawler) deploy onto a lightweight Kubernetes cluster, while self‑hosted GitHub Actions runners handle CI. Every layer is defined, built, and operated through Terraform, Packer, and ArgoCD.