I build and operate cloud platforms that help engineering teams deploy, recover, and maintain production services with confidence. I bring 10+ years of experience in IT, hands-on infrastructure ownership, and technical leadership working directly with CTOs.
Based in Caracas, Venezuela (UTC−4).
- Platform ownership: Built an Azure production platform from the ground up with AKS and infrastructure automation, covering delivery workflows and ongoing operations.
- Developer enablement: Designed component-specific CI/CD pipelines and FluxCD GitOps workflows to make deployments repeatable and changes reviewable. Previous experience includes Jenkins delivery workflows on AWS.
- Architecture and migration leadership: Led cloud migration initiatives directly with CTOs and maintained technical continuity through CTO transitions.
- Infrastructure as Code: Used Terraform to reproduce an enterprise architecture for a Badgr deployment, translating infrastructure requirements into repeatable configuration.
- AI infrastructure: Deployed and operated GPU-backed LLM services on Kubernetes.
- Security and operations: Implemented infrastructure controls supporting SOC 2 examinations, connecting platform configuration with operational practices and evidence.
A reproducible PostgreSQL recovery lab built around the question: can a backup actually restore usable data after a failure? It implements data-loss, service-loss, and corrupted-backup scenarios using disposable Docker containers and synthetic data.
- Logical backups with
pg_dump, clean-room restores withpg_restore, and SHA-256 archive integrity checks. - Exact baseline comparisons, an expected post-backup data-loss check, and a database write probe.
- Database-level recovery-time and recovery-point-age measurements against configurable targets, with JSON evidence reports.
- Documented operator runbook, recovery tradeoffs, and explicit validation limits.
Demonstrates: recovery engineering · PostgreSQL · Python · Docker · RTO/RPO reasoning · operational documentation. This is a local reference lab; production failover and end-to-end service recovery are outside its implemented scope. See the repository for validation status.
A multi-cloud platform for infrastructure control analysis and Terraform bundle generation. Combines a Go API and React dashboard with AWS, Azure, and GCP modules, control mappings, scan history, and reviewable remediation proposals.
Demonstrates: Terraform · cloud security · SOC 2 control mapping · Go · multi-cloud architecture.
| Project | Engineering focus |
|---|---|
| GitOps Developer Platform | Kubernetes service scaffolding, optional Flux resources, and Git-based rollback guidance. |
| Infrastructure Drift Monitor | Python CLI reporting recorded drift and proposed changes from Terraform JSON plans. |
| LLM Inference Platform | Kubernetes manifests for GPU inference and compute-cost estimates from supplied rates. |
| Cloud Reliability Lab | SQLite backup and restore, request-based SLO reports, and recovery runbooks. |
| AI DevOps Prompts | Reusable prompts for CI/CD reviews, testing, and troubleshooting, with a worked example and evaluation criteria. |
These are independent reference implementations. Each repository documents its scope, design decisions, and validation limits.
Platform: Kubernetes · AKS · Docker · Helm · Linux
Infrastructure automation: Terraform · Ansible · Python · Bash
Delivery: GitHub Actions · FluxCD · Jenkins · ArgoCD · Git
Cloud: Microsoft Azure · AWS · Google Cloud · Oracle Cloud Infrastructure
Data and workloads: PostgreSQL · Redis · GPU-backed LLM inference
- GeoVZ: A web geography game focused on Venezuela, combining real-world imagery, location discovery, and competitive play. Source code is private.
- Planea: An AI-assisted local discovery and planning platform for businesses, services, and events in Caracas. Source code is private.
I'm interested in teams that need an engineer who can own platform architecture, improve delivery workflows, and carry changes through production operations.
- Location: Caracas, Venezuela · UTC−4, aligned with U.S. East Coast working hours.
- Opportunities: Remote Senior Platform Engineer roles and long-term international contractor engagements with U.S. companies.
- Languages: Spanish · English · Portuguese.
- Contact: LinkedIn · juanmdsousa@gmail.com.
Professional contributions from the past two years also appear under my dedicated work account, @jdesousa-bc.