LEVI AI just now
Levi Lugato
Staff / Lead DevOps & Platform Engineer
I design the platforms behind the product: resilient cloud foundations, safe delivery paths, useful observability, and automation teams can trust.
levi@platform:~$ whoami
levi@platform:~$ cat current-focus.txt
LEVI AI context loaded
Agentic Platform Engineer builder
Try a command to navigate this portfolio.
scroll to inspect the context ↓
01
levi@platform:~$ cat about.md
About me
LEVI AI personal context
I was born in Viradouro, a small city in the countryside of São Paulo, Brazil. I got into IT in 2012 by working close to the hardware: running cable, configuring firewalls, looking after on-premises servers, and supporting Windows, Linux, and the network between them.
That hands-on start made me curious about how the whole system fit together, not only the piece directly in front of me. When something failed, I wanted to follow the problem all the way through: from the physical connection and operating system to the service and the person trying to use it.
I began in traditional infrastructure, supporting Linux, Windows, FreeBSD, networks, firewalls, virtualization, web servers, and relational databases. As my work moved into automation and cloud, I started provisioning AWS and Azure environments with Terraform, Ansible, and Packer, building CI/CD pipelines, containerizing services, and writing Python and shell tooling to replace manual operations. That led into SRE, Kubernetes, GitOps, observability, incident response, and the daily work of keeping production systems reliable across cloud and hybrid environments for international teams.
Today I work as a Staff / Lead Platform Engineer. I design and operate multi-account, multi-region AWS platforms on EKS and ECS Fargate, build reusable infrastructure with Terraform, AWS CDK, and CloudFormation, and create delivery paths with GitHub Actions, GitLab CI, Azure Pipelines, and AWS CodePipeline. My work also covers OpenTelemetry, Prometheus, Grafana, CloudWatch, Datadog, SLOs, developer self-service, and the operational standards teams need to run services safely. I have worked fully remotely since 2019 with teams across Latin America, the United States, Canada, Europe, and Asia. I still like getting close enough to a problem to understand it, then automating the repetitive parts so the next person has less to fight with.
trajectory.log
On-premises infrastructure, cabling, networks, firewalls, servers, Linux and Windows.
Infrastructure as Code, AWS and Azure, containers, delivery pipelines, and architecture.
Kubernetes, GitOps, observability, SLOs, and production systems across global teams.
Multi-account, multi-region AWS platforms and self-service paths for engineering teams.
02
levi@platform:~$ ls -la work/
What I build
drwxr-xr-xplatform-foundations/01
Cloud platforms that give teams a paved road.
Reusable infrastructure, sensible guardrails, and self-service patterns across multi-account and multi-region AWS.
- Terraform
- AWS CDK
- AWS Lambda
- Kubernetes
- EKS
- ECS Fargate
- Ansible
- Packer
drwxr-xr-xdelivery-systems/02
Delivery paths built for speed and recovery.
Traceable CI/CD, GitOps, blue/green and canary releases, with rollback close at hand.
- GitHub Actions
- GitLab CI
- AWS CodePipeline
- Azure Pipelines
- CircleCI
- Helm
- Blue/green
- Canary
drwxr-xr-xautomation-scripts/03
Small programs that remove repetitive work.
Serverless functions, operational tooling, and repeatable automations built in Python, Bash, and JavaScript for scheduled jobs and everyday engineering workflows.
- Python
- Bash / Shell
- JavaScript
- TypeScript
- Node.js
- AWS Lambda
- EventBridge
drwxr-xr-xreliability/04
Signals that explain the system, not just its symptoms.
Telemetry, SLOs, incident practices, and operational standards that turn production noise into useful context.
- OpenTelemetry
- Prometheus
- Grafana
- AWS CloudWatch
- New Relic
- Datadog
- Sumo Logic
03
levi@platform:~$ jq . resume.json
Experience, parsed
{
"role": "Staff / Lead Platform Engineer",
"experience": "13+ years",
"cloud": ["AWS", "Azure", "GCP", "OCI"],
"platforms": ["Kubernetes", "EKS", "ECS"],
"iac": ["Terraform", "AWS CDK", "CloudFormation"],
"languages": ["TypeScript", "Python", "Ruby", "Bash / Shell"],
"systems": ["Linux", "FreeBSD", "Windows Server"],
"runtimes": ["Node.js", "Ruby on Rails", "Rust", "Elixir"],
"data_stack": ["Trino", "Presto", "Apache Hive", "Apache Spark", "Apache Iceberg", "Apache Superset", "AWS Glue", "Amazon Athena", "Amazon Redshift", "S3 Data Lake"],
"certified": ["CKA", "CKAD"],
"remote_since": 2019
}recent_history --limit 4
Arionkoder / iSeatzLead / Staff Platform Engineer
LumenaltaSenior DevOps Engineer
NavisiteSenior DevOps Engineer
Codeminer42Site Reliability Engineer
Earlier: Hapvida · Beblue · Magazine Luiza / LuizaLabs · Linx
04
levi@platform:~$ ls -lt blog/
Field notes
Reducing Token Usage in Codex with Model Routing, RTK, and Code Graphs
How to divide read-heavy exploration, mechanical work, and critical decisions across models—reducing the main agent’s context without outsourcing architecture, security, or accountability.
cat article.md ↵05
levi@platform:~$ ./contact.sh
Let’s solve a hard systems problem.
LEVI AI ready
Tell me what is unreliable, repetitive, slow, or difficult to operate. That is usually where the useful work begins.