Cloud product architecture
Target architecture for SaaS, PaaS, ERPNext, and service operations - scoped around ownership, rollout, and day-2 operations.
- SaaS • PaaS • ERPNext
- Tenant models • onboarding
- AWS • Azure • GCP • IONOS
Cloud architecture for SaaS, ERP, and managed services.
I design tenant models, automation, security boundaries, and operating models behind cloud products your team can launch, run, or turn into a managed service.
What you get
Reproducible infrastructure, not tribal knowledge
Terraform and GitOps replace manual setup with reviewed, repeatable changes in version control.
- Terraform • Providers • Ansible
- ArgoCD • Flux • GitOps
- GitLab • GitHub Actions • CI/CD
Operating model and handover
Decision records, runbooks, monitoring paths, and team enablement so the service can be run internally or offered as a managed product.
- Workshops • Pairing • Coaching
- Runbooks • Support boundaries
- Architecture and stakeholder advisory
AI and RAG service architecture
RAG, MCP, and agent systems designed with tenancy, observability, access control, cost awareness, and hybrid hosting options.
- AI inference • RAG • Agents
- LLM orchestration • MCP
- Hybrid cloud • Multi-tenancy • MLOps
Engagement Models
Work with me on architecture, assessment, or scoped implementation support for SaaS, ERP, AI/RAG, or managed-service foundations. Remote-first from Germany. You own the decisions, code, infrastructure definitions, docs, and operating patterns.
Cloud Architecture & Handover
Design • Automate • Transfer
- Scope: Define and implement the architecture for a cloud product, ERPNext service, or managed-service foundation with the automation to operate it
- Stack: Terraform, Kubernetes, GitOps, Ansible, observability
- Deliverables: Target architecture, tenant model, infrastructure baseline, operating workflows, pipelines, runbooks, dashboards, handover
- Typical timeline: 3-6 months, depending on scope
Assessment & Advisory
Review • Roadmap • Decisions
- Scope: Review cloud architecture, delivery workflows, operational readiness, and the roadmap for safer implementation
- Format: Workshops, pairing, async review, architecture sessions
- Deliverables: Findings, decision records, reference patterns, prioritized next steps
- Typical timeline: 1-3 weeks for assessments; retainer for ongoing advisory
AI Product Infrastructure
RAG • MCP • Agents
- Scope: Move AI/RAG workflows from prototype to operated product: tenant model, data paths, secure access, observability, and cost controls
- Stack: Python/Go, LLM orchestration, vector search, cloud infrastructure, CI/CD
- Deliverables: Service architecture, APIs, deployment path, monitoring, runbooks
- Typical timeline: 4-12 weeks, depending on product scope
How it works: Scoping call -> written scope -> weekly working sessions -> handover. You own code, docs, and IP.
What working with me looks like
- Discover: Map the product goal, constraints, delivery path, and future ownership model.
- Architect: Turn the target state into clear decisions, trade-offs, and implementation slices.
- Validate: Prove the design through reference implementation, automation, or delivery slices.
- Enable: Pair with the team, document the system, and hand over the operating model.
Reference Implementations
The same architecture patterns across production systems: a Kubernetes workflow engine, a local-first AI agent runtime, ERPNext service automation, and multi-tenant SaaS operations.
ERPNext Provisioning System: Designed and implemented a multi-tenant SaaS service for operating isolated ERPNext instances on IONOS Cloud. The system includes a custom Kubernetes operator in Go, a self-service administration panel, and automated day-2 operations through Terraform and GitOps.
Selected Architecture Work & Clients
- Multi-tenant ERP service
- Direktiv.io — K8s workflow engine
- Contenox — AI agent runtime
- Telekom Innovation Lab
- University of Bremen
Common Questions
Can you design a multi-tenant SaaS or ERP service?
Yes. I design tenant models, control workflows, Kubernetes operators where useful, IaC, GitOps delivery, monitoring, and the operating model needed to run it.
Do you also handle implementation?
Yes, when it helps prove or deliver the architecture. The center of the engagement is the target design and operating model; implementation is scoped around the parts that unblock delivery and handover.
Can this support a future managed-service offer?
Yes. I can shape the architecture, automation, monitoring, support boundaries, and handover so a SaaS, ERP, or AI/RAG service can later become a managed offer under the right brand.
What does "enablement" mean in practice?
Workshops, pairing, runbooks and best practices so your team owns the service after handover. The goal is a self-sufficient team, not a dependency on me.
Which clouds and tools?
AWS, Azure, GCP and IONOS; Terraform (incl. custom providers), Ansible, Kubernetes, ArgoCD/Flux, GitLab and GitHub Actions; Prometheus, Grafana, OpenTelemetry. Secure defaults throughout: secrets management, least-privilege, scanning, review gates.
Availability & engagement?
Remote-first from Germany, with occasional on-site when it helps delivery. Best fit: SaaS/PaaS architecture, ERPNext service operations, AI/RAG product infrastructure, Kubernetes automation, assessments, and advisory retainers where implementation and handover matter. Contract details on request.
Get In Touch
Let's discuss the architecture
Send the product or service goal, current constraints, and what you want to operate now or offer later. I will reply with whether architecture, assessment, or scoped implementation support is the right next step.
Hamburg, Germany