Problems solved, step by step.

Selected work from 10+ years in cloud and infrastructure. Clients and employers stay anonymous. The problem, the approach and the result are what matter here.

1. 01 / Product company AWS · Google Cloud · Terraform · Ansible

   Google Cloud to AWS, zero downtime

   1. Challenge

      Every production and staging workload had to move from Google Cloud to AWS while the product stayed live.

   2. Approach

      Designed the target environment on AWS, codified it in Terraform and Ansible, and cut over one environment at a time.

   3. Result

      All workloads on AWS with zero downtime during the move.

2. 02 / Product company SonarQube · Wazuh · Vaultwarden · Jenkins · GitHub Actions

   A self-hosted DevSecOps platform

   1. Challenge

      Code quality, threat detection and credentials were handled differently by every team.

   2. Approach

      Rolled out SonarQube, Wazuh SIEM and Vaultwarden, wired into CI/CD on Jenkins and GitHub Actions.

   3. Result

      One shared stack for code quality, threat detection and secrets, and repeatable deployments.

3. 03 / Enterprise clients Security Command Center · Cloud Armor · WAF · SIEM · Zero trust

   Security hardening for PCI DSS

   1. Challenge

      Cloud platforms had to meet PCI DSS and local regulatory requirements.

   2. Approach

      Hardened the security posture with SCC, SIEM, SOAR, WAF, Cloud Armor and zero trust access.

   3. Result

      PCI DSS and local regulatory compliance achieved.

4. 04 / Enterprise clients Terraform · Ansible · CloudFormation · Cloud Build

   Multi-cloud automation

   1. Challenge

      Public, private, hybrid and multi-cloud estates on GCP, AWS and Azure were being built by hand.

   2. Approach

      Automated network, compute and virtualization layers with Terraform, Ansible, CloudFormation and Deployment Manager.

   3. Result

      Infrastructure that can be rebuilt from code, with each migration planned, carried out and validated.

5. 05 / Data team Python · Gemini · OpenAI · OCR

   An AI-powered public data lake

   1. Challenge

      Indonesian public datasets were scattered across PDFs and portals and collected by hand.

   2. Approach

      Built Python scrapers with Gemini and OpenAI for PDF extraction and OCR, feeding a structured data lake.

   3. Result

      Manual collection replaced, with near-real-time access to the datasets.

6. 06 / Product company Terraform · Ansible

   Environments without drift

   1. Challenge

      Development, staging and production had drifted apart through manual changes.

   2. Approach

      Standardized provisioning in Terraform and Ansible, so every environment is built the same way.

   3. Result

      Less manual configuration and no more environment drift.