Azure · AWS · Kubernetes · AI Platforms. I design and scale the cloud infrastructure and developer experience behind mission-critical microservices and next-generation AI platforms — then map it so anyone can understand it.
Numbers describe Ultraplatform — the reference platform I model in the 3D map, kept in sync by its live engine. My real-world career impact lives on the CV.
I'm a Senior (AI) Platform Engineer and Cloud Architect (AWS Certified Solutions Architect – Associate · CKA) working where large-scale platform engineering meets the AI revolution — designing, scaling and optimizing the cloud infrastructure behind mission-critical microservices and a company's next-generation AI platforms.
My focus is Developer Experience (DevEx): abstracting Azure and AKS so product teams ship safely without being infrastructure experts — golden paths, self-service, policy-as-code guardrails and observability by default.
Over fifteen years I've gone from full-stack product engineering (Klarna, Berlin) to multi-cloud infrastructure (AWS/EKS & Azure/AKS) to designing platforms end-to-end. I care about cost-aware architecture, test-driven infrastructure and making complex systems legible — which is what Ultraplatform, my interactive map, is for.
Off the clock I'm a happy father of three. I love learning, good conversations and weekends in the mountains with my family. My wife is from El Salvador — a country I love for its nature and its people. I value teamwork, and roles where I can help others do their best work.
Patterns from the platform I model in the map — each one a problem teams kept hitting, the design that turned it into something self-service, and a link to the proof.
terraform apply away, caught only if someone noticed in review.
From full-stack product work to multi-cloud platforms — the throughline is reliability, cost-awareness and developer experience.
Design, scale and optimize the cloud infrastructure powering mission-critical microservices and the company's next-generation AI platforms. Lead Developer Experience: abstracting Azure/AKS complexity behind golden paths, self-service and policy-as-code. Explore the platform I build →
Built and operated the Azure/AKS developer platform: reusable Terraform modules, GitLab CI components, multi-region clusters, and the guardrails that keep tenant teams safe by default.
Designed clean, maintainable architectures for enterprise applications (incl. OutSystems), focusing on separation of concerns and long-term maintainability.
Directed the Business Analytics master's programme — curriculum and mentoring bridging data, analytics and engineering practice.
Migrated infrastructure to AWS (EKS) and Azure (AKS): Packer + Ansible base images, Terraform for the infrastructure, and an own Test-Driven Infrastructure (TDI) framework in Go. Adopted the AWS Well-Architected Framework.
Google Pay API integration and a merchant onboarding API (OAuth2). Built an own TDD testing framework; microservices in Java 11 / Spring Boot with React/Redux front-ends, Kafka, and Jenkins CI/CD.
Full-stack PHP/JS (Zend, MySQL, jQuery plugins), database engineering for a social platform, an iOS game (cocos2d / Objective-C), and SCORM e-learning courseware.
Platform, cloud and delivery — plus the product-engineering background that makes the platform actually usable.
It runs entirely in your browser — nothing is sent anywhere — and shows which of my core areas your role touches, each with a link to proof.
Not a logo wall — every tool here runs in the platform today and has an article explaining the job it does and what I'd tell you before adopting it.
The packaging grammar — one chart library, consumed everywhere.
Read →GitOps delivery — the cluster state is a repo, not a memory.
Read →In-cluster policy: validate, mutate, generate — safe namespaces by default.
Read →Key Vault → Kubernetes secrets, without humans in the loop.
Read →Config and secret changes actually reach running pods.
Read →DNS records follow the ingresses that need them.
Read →Scale on real signals — queues and events, including to zero.
Read →One instrumentation standard every tenant inherits.
Read →Debug against the real cluster from the laptop.
Read →Just-in-time, right-sized nodes — provisioned in seconds, consolidated when idle.
Read →Dashboards, alerts and Explore as code — one pane over Mimir & Loki.
Read →The developer portal — software catalog, golden-path scaffolder and TechDocs.
Read →Images without Dockerfiles — one hardened builder, no per-repo drift.
Read →Certificates as a controller — TLS issued and renewed automatically.
Read →Continuous vulnerability scanning — CVEs found after deploy, not just at build.
Read →One dashboard for every policy engine's reports — Kyverno and Trivy in one pane.
Read →Cost allocation per tenant and namespace — the bill made visible.
Read →Infrastructure can't change unless policy says it may — a plan-time safety gate.
Read →Runtime security — catching what admission policy and image scans can't see.
Read →Chaos as a reviewed experiment — never ambient, always opt-in.
Read →Decouple the deploy from the release — vendor-neutral feature flags.
Read →The eBPF CNI, chosen at cluster creation — L3-L7 policy and Hubble visibility.
Read →Sidecar building blocks — pub/sub, state, durable workflow, one API.
Read →One shared gateway, many tenant routes — SSO by sidecar, not by app code.
Read →The infrastructure I run isn't the only thing I build — I write the tools that make the day-to-day loop fast and safe.
Run the reusable pipeline locally and diff it against remote before you push.
See in the map →Apply and clean up Terraform overrides locally, faithful to the remote steps.
See in the map →A tmux time tracker that follows your ticket and feeds standups automatically.
See in the map →Mirror a personal Notion workspace as a local markdown vault — both ways.
View on GitHub →Daily drivers — not mine, just essential: kindskaffoldtelepresence k9ssterndive gitlab-ci-local
Grafana Alloy configures with River — I built its missing editor experience end to end.
Open to conversations about platform engineering, AI platforms, and developer experience.