Mateus – Kubernetes, Terraform, Ansible

An AI cluster orchestration platform for managing multiple Slurm and Kubernetes environments across several cloud providers. It provides fleet health monitoring and an alerting strategy designed to help maintain SLA targets.

• Created and managed CI/CD pipelines in GitLab CI;
• Used Terraform to deploy Kubernetes and Slurm clusters with hundreds of GB of NVIDIA GPUs (H100, H200, GB200, GB300);
• Deployed Kubernetes and Slurm clusters in AWS, GCP, Azure, Nebius, and Oracle;
• Troubleshot Linux environments;
• Managed Slurm job operations;
• Implemented TailScale, enabling users to access the Slurm Login Node securely;
• Implemented ArgoCD using Kustomize for bootstrap;
• Used Pulumi (with Python) to deploy AWS infrastructure for internal tools;
• Integrated Grafana with Terraform to improve observability by creating dashboards for system metrics and Slurm fleet health;
• Used Alloy to manage data pipelines from Kubernetes to Grafana Cloud;
• Created a maintenance strategy to reduce false positive alerts through automation;
• Implemented a Slurm exporter written in Golang to enrich observability across multiple cloud providers.

A fitness app platform focused on global delivery and cost-optimized infrastructure across Azure services. It relies on a CDN strategy and cloud architecture improvements to support user experience at scale while keeping infrastructure costs under control.

• Managed multiple subscriptions with Terraform;
• Modularized Terraform code for deployment across multiple environments;
• Implemented a CI/CD pipeline in Azure DevOps to deploy Terraform and application code;
• Improved Datadog monitoring capabilities through code (IaC);
• Reduced tenant costs by 25% by optimizing AKS and Azure Functions;
• Reduced database costs by 15% for Azure Cosmos DB by optimizing RU (Request Units);
• Improved user experience by creating a CDN cache layer in Azure Front Door;
• Used APIM to version developer APIs, integrating them with the Azure Front Door solution.

An infrastructure and delivery platform for an international banking environment, focused on Terraform- and Bicep-based provisioning with Azure DevOps as the CI/CD tool. It supported secure, reliable, and governed infrastructure delivery across Azure and GCP.

• Architected, implemented, and documented a disaster recovery strategy to ensure data security and reliability;
• Produced post-mortem documentation to improve operational processes and systems;
• Applied policies with infrastructure as code and CI/CD to streamline governance and automate compliance;
• Applied security practices with SonarQube, Fortify, infrastructure as code, and CI/CD;
• Facilitated agile, scalable, and efficient infrastructure management in Azure and GCP using Terraform and Bicep;
• Implemented GitOps using ArgoCD on GKE;
• Enhanced visibility and control over software deployments using CI/CD techniques and observability;
• Enhanced security with automation and CI/CD techniques using Python scripts;
• Used Azure DevOps to boost team productivity and support high-quality software releases.