INDEPENDENT ENGINEER. PRODUCTION MINDSET.
I’m Stephano. I design platforms, automate delivery, and keep production running — from cloud infrastructure to AI systems.
Platform Engineering / DevOps / SRE / MLOps
ENGINEERING EXPERIENCE AT
01 /WHAT I BRING
Cloud architecture, deployment automation, and the work of keeping services healthy. I build and operate the platform, whether it runs your core product or your AI.
A foundation your team can build on.
Build a Kubernetes platform with repeatable environments, self-service delivery, and infrastructure your team can change with confidence.
Ship with confidence. Sleep better.
Release pipelines, service-level objectives, observability, and incident response. I build the delivery workflows and operating practices that keep software shipping and services running.
From promising model to production AI.
Put trained models and AI workflows on infrastructure you can operate. Model serving, GPU capacity, release pipelines, and the controls around AI agents.
DEVOPS, SRE & PRODUCTION AI
Production AI depends on the same disciplines as any dependable service: sound architecture, repeatable releases, capacity planning, and incident response. I bring that DevOps and SRE experience to model serving, GPU infrastructure, and AI tooling.
See the ai-coustics workCode, infrastructure, model artifacts.
Tests, security checks, performance.
Canary traffic. Defined rollback.
SLOs, service health, inference signals.
02 /SELECTED WORK
From AI startups to national infrastructure. A few places where the engineering has to hold up.
Real-time speech enhancement leaves little room for slow inference or interrupted service. I built and operated the platform behind it, spanning cloud architecture, Kubernetes, release engineering, and reliability.
I kept production serving through releases and infrastructure changes without downtime. I designed GitHub Actions workflows for reviewed, repeatable delivery, built an AI harness and custom Kubernetes tooling, and handled GPU capacity, model rollout, and observability.
Model and service releases without interrupting production, backed by controlled rollouts, clear rollback paths, and visibility into the running systems.
As a core engineer in a small team, I architected and implemented Volkswagen’s Kubernetes and container orchestration platform across AWS and Azure, working through Storm Reply.
My work covered the entire platform: cloud foundations, cluster architecture, deployment automation, and operations across dozens of clusters and 500–1,000 autoscaled nodes. I also owned SRE, from SLOs and on-call to incident response, critical etcd fixes, and upstream contributions to WeaveNet CNI.
Infrastructure supporting 200+ applications and more than 3 million connected vehicles.
Germany’s national air navigation service provider. A cloud and Kubernetes platform used by teams running flight control systems, airport tower systems, and operations.
I delivered automated infrastructure across private-cloud and data-center environments: provisioning and managing VMs, building and bootstrapping Kubernetes clusters, and connecting their lifecycle through Terraform, Ansible, and GitLab CI. Changes followed reviewed pipelines, clear environment boundaries, and controlled promotion into production.
Infrastructure delivery under regulated-industry requirements, with change control and a defined promotion path.
Selected experience across freelance and employed roles. Volkswagen engagement through Storm Reply.
BEYOND INFRASTRUCTURE
My security scanner for AI agent configurations and CI/CD pipelines. Written in Go, with 70+ rules covering agent configs, MCP, prompt injection, and the software supply chain.
03 /THE ENGINEER BEHIND THE PLATFORM
Building systems since 2006. Still hands-on.
I started with operating systems and network appliances. Since then, I’ve written backend services, built Kubernetes platforms, and taken the call when production broke. That experience shapes how I build cloud platforms and production AI systems.
I’m an independent engineer. I take responsibility for the architecture, the code, and the day-to-day operation — and make sure your team can own what we build.
Some problems call for Go or Python. Others need a better deployment workflow, a clearer SLO, or less infrastructure. I start with the problem.
Building systems since
Years of production ownership
Vehicles on the VW platform
Continents in the Signavio platform
04 /WORKING TOGETHER
I work in your codebase, alongside your engineers, and take changes through to production. Architecture decisions and operational responsibility stay connected.
We map the system, the bottlenecks, and what success looks like. You get a clear scope and an honest technical assessment.
Changes are reviewed, tested, and rolled out in steps your team can follow. You see the work in your codebase and the results in your environment.
Tested systems, useful documentation, and knowledge that stays with your team. Continued support when you need it.
Available for remote engagements worldwide.
Yes. I’m available for remote engagements worldwide. Based in Berlin, I agree on practical meeting windows and collaboration hours with each team. English is my main working language; I also speak fluent Portuguese and Spanish, and intermediate German.
Yes. My focus is the platform around your model: packaging and versioning, inference serving, GPU scheduling, deployment pipelines, progressive rollout, and observability. I work with your ML engineers so the model becomes a service the rest of the business can depend on.
It can be a focused infrastructure project or an ongoing embedded engagement. We agree on scope, ownership, availability, and commercial terms after discussing your needs. I work directly with your team and write the code, not just the recommendations.
Email contact@szfs.eu with a few lines about your platform, your biggest bottleneck, and your timeline. We’ll discuss the situation, check whether I’m the right fit, and define a useful first step.
GOOD SYSTEMS START WITH A CONVERSATION
A platform to build. Releases to untangle. Production to stabilize. Or an AI system ready for real traffic. Tell me what you’re working on.
Start a conversation