Skip to main content
/
Work / 20+ years of experience

Technical Lead and Solutions Architect.

I get to the bottom of a problem and solve it where it belongs. At Douglas that took me from the React frontend of the POS software into the CRM backend, where I designed one unified API so business logic no longer had to be duplicated across several frontends.

Discuss a projectDownload project dossier

Availability

Available · 2–5 days/week · Munich and remote DACH · German and English

Where I tend to help

Architecture modernisation · Delivery reliability · Governed AI workflows

Profile picture of Martin Trenker, independent software consultant

Martin Trenker

Technical Lead and Solutions Architect

What drives me

Improving digital systems and products, and with them the daily lives of the people who work with and use them. That starts with understanding their problems and finding the best way to solve them. I picked that up early: as a teenager I rented my first server and helped people with web development on IRC.

Core technologies

dotnetcsharpgraphqlresttypescriptnode.jsnestjsexpressreact

Selected projects

Three recent projects: one CRM API instead of business logic duplicated across frontends, a Next.js platform handed over from an agency to a new in-house team during active delivery, and an AngularJS workflow rebuilt in React on an explicit state machine.

Counts and periods from the project records are marked as measured facts. What I observed but did not measure is marked as an observation.

React Expert → Technical Lead → Solutions Architect

Douglas GmbH

Jan 2024 – Jul 2026

Mandate

Modernise the POS/CRM platform from its React frontend to backend services in TypeScript/Node.js and C#/.NET, unify the CRM API landscape, and introduce governed AI tooling.

Role progression

  1. 01

    React Expert

    Joined to modernise the POS software used in more than 1,200 stores across 14 European countries.

  2. 02

    Technical Lead

    Led the move from legacy Angular to maintainable React and shipped the business-critical CRM integration in under a year.

  3. 03

    Solutions Architect

    Took on broader scope across TypeScript/Node.js and C#/.NET backend services to unify the CRM API landscape and introduce governed AI tooling for daily delivery work.

Why a separate C#/.NET API

Extending the existing NestJS/TypeScript API was one of a few options. Internally, though, there was a move towards C# on the backend, because it was the best understood and supported setup. The existing API was also not built from the ground up for the resilience and scaling the unified API required; adding another controller to it would have meant shoehorning those requirements in. I continued to maintain the production TypeScript API.

Outcomes and evidence

  • Measured fact

    software runs on every cash register in more than 1,200 stores across 14 European countries

  • Measured fact

    rewrote the CRM/POS integration in under a year with clearer architecture, more test coverage, and updated documentation

  • Unmeasured observation

    moved from after-hours deployment windows toward safer deployments during business hours through better error handling, observability, and recovery paths

  • Unmeasured observation

    built an in-house shortlink and QR service in Next.js + .NET in weeks, from request to production, replacing a third-party SaaS that no longer met business needs

Personal ownership

  • +extended the solution into C#/.NET and personally designed and shipped the unified API, applying architecture patterns across stacks and using AI tooling as part of a reviewable engineering workflow
  • +replaced static code-level tenant configuration with dynamic runtime config served by the API — config changes no longer require a release
  • +developed and maintained a production CRM backend in TypeScript/Node.js with NestJS and Express, covering API endpoints, configuration, and external service integrations
  • +built a Node.js/TypeScript agent CLI giving AI agents governed access to Jira, Confluence, Azure DevOps, and internal tooling, with human review of all output

Team delivery and contribution

  • +onboarded 1 senior and 2 junior developers; all productive contributors within weeks
  • +introduced "Full Stack Fridays", team-driven learning sessions covering Figma prototyping, GraphQL, C#/.NET, and more

Solutions Architect

Oetker Digital GmbH

Sept 2022 – May 2023

Situation

A Next.js platform was moving from an external agency to a newly formed in-house team while active delivery continued.

Mandate

Audit the codebase, consolidate two products, and establish the new team without stopping ongoing product work.

Personal ownership

  • +audited the codebase and improved quality across the stack: Next.js, TypeScript, GraphQL, Apollo, Cypress, Storybook, Tailwind
  • +consolidated two separate web products into a single codebase without disrupting active development

Outcomes and evidence

  • Unmeasured observation

    consolidated two separate web products into a single codebase without disrupting active development

Team delivery and contribution

  • +onboarded four new hires in parallel while keeping delivery on track
  • +introduced lightweight agile practices that the team adopted

Technical Architect

Fielmann AG

Mar 2021 – Dec 2021

Situation

An AngularJS optometry workflow for in-store staff had accumulated complex, ad-hoc state logic.

Mandate

Rebuild the workflow in TypeScript and React with an explicit state-machine architecture and transfer the knowledge into the team.

Personal ownership

  • +state-machine driven multi-step form architecture built to make changes safer
  • +custom on-screen keyboards tailored to in-store workflows

Outcomes and evidence

  • Unmeasured observation

    state-machine driven multi-step form architecture built to make changes safer

Team delivery and contribution

  • +daily pair-programming as the main knowledge transfer strategy

Current hands-on systems

Products, tooling and the platform they run on, which I build, operate and pay for myself. Here I own every layer: data model, application code, deployment, observability and recovery.

EU-sovereign Kubernetes platform

since 2025

After five years of running my own projects serverless on AWS (CDK, Lambda, Cognito), I migrated everything in 2025 to a self-operated Kubernetes platform on European infrastructure — and have run it in production since.

What I own
I provision the cluster and own deployment, observability, data services, secrets, upgrades, and recovery.
Client relevance
Direct operating experience with GitOps, Kubernetes, and production observability informs architecture and delivery decisions for client systems.
Operational responsibility
Five production applications run on the platform with PostgreSQL pooling, alerts, dashboards, logs, and metrics.
  • +Terraform-provisioned cluster on Hetzner with Cilium CNI
  • +GitOps deployments via ArgoCD; secrets encrypted in Git with Sealed Secrets
  • +full observability stack: Prometheus, Grafana, Loki
  • +PostgreSQL via CloudNativePG with pooling, alerts, and dashboards
  • +runs five production applications
kubernetesterraformargocdciliumsealed-secretsprometheusgrafanalokipostgresql

yamiat.com

since 2020

A production side project in continuous development since 2020. From 2020 to 2025 it ran serverless on AWS using CDK, Node.js Lambdas, AppSync, Cognito, EventBridge, and a single-table DynamoDB model.

What I own
I owned the full AWS architecture, infrastructure as code, data model, application code, and self-mutating deployment pipeline.
Client relevance
Hands-on production experience with serverless APIs, SQS/SNS messaging between services, S3 static asset hosting via CloudFront, and DynamoDB access patterns.
Operational responsibility
Operated on AWS for five years, then migrated to a self-operated Kubernetes platform in 2025; continuously developed in production since 2020.
awsaws-cdklambdanode.jsappsyncdynamodbcognitoeventbridgesqssnss3cloudfrontcodepipelinenext.jsreacttypescriptgraphqldotnetpostgresql

Agent tooling & MCP

since 2026

The workflow behind my daily AI usage: a custom MCP server, skills and extensions for coding agents, and deterministic tooling around AI output — built and used daily, not demoed.

What I own
I build and maintain the MCP server, agent skills, extensions, and deterministic checks around generated work.
Client relevance
The system makes AI-assisted delivery bounded, inspectable, and reviewable instead of treating model output as an unchecked result.
Operational responsibility
Used in daily engineering work with human review before changes are accepted.

Reviewable workflow example

Reviewable example: issue-scoped agent → isolated worktree → deterministic test and build gates → human-reviewed pull request.

typescriptmcpai-agentsnode.js

Project timeline

  1. Douglas GmbHSolutions ArchitectContract extended with broader scope across the CRM backend landscape. Extended the existing production TypeScript/Node.js API built with NestJS and Express, designed and shipped a unified C#/.NET API (REST + GraphQL), and introduced governed AI tooling for daily delivery work.
  2. Douglas GmbHTechnical LeadEngaged as React expert to modernise the POS software used in more than 1,200 Douglas stores across 14 European countries. Led the transition from a legacy Angular implementation to a maintainable React architecture and shipped the business-critical CRM integration in under a year.
  3. Oetker Digital GmbHSolutions ArchitectManaged the handover of a Next.js platform from an external agency to a newly formed in-house team, audited the codebase, merged two products, and got the team up to speed.
  4. UXMA GmbH & Co. KGTechnical LeadLed a three-person team building a data-heavy internal web app in TypeScript and React, with a strong focus on mentoring two junior developers.
  5. Fielmann AGTechnical ArchitectRewrote a complex AngularJS optometry workflow app in TypeScript/React, replacing ad-hoc state logic with XState-driven form architecture for in-store staff.
  6. Interhyp AGEngineering ConsultantBuilt a real-time communication micro-frontend handling VoIP calls between customers and mortgage advisors, using React, Redux Toolkit and WebRTC.
  7. ProSiebenSat.1 Digital GmbHFrontend EngineerBuilt and maintained React components powering high-traffic show websites for ProSieben and Sat.1 formats, with GraphQL APIs and analytics tracking throughout.
  8. InstaMotion Retail GmbHFrontend EngineerBuilt a React + GraphQL web magazine for Germany's largest online car dealer, connected to a headless CMS.
  9. Fineway AGFrontend EngineerBuilt components for an online travel magazine using React, TypeScript and GraphQL, backed by Contentful as a headless CMS.
  10. Lufthansa AGFrontend EngineerDeveloped React components for Lufthansa's digital touchpoints, integrated into Adobe Experience Manager.

Certifications: AWS Certified Developer – Associate & AWS Certified Cloud Practitioner (until 2023).

AvailabilityAvailable · 2–5 days/week · Munich and remote DACH · German and English

If this sounds close to what your team needs, let’s discuss the project.

Send a short note about the team, the transformation pressure, and what breaks between intent and execution. If I can help, I will. If not, I will say so honestly.

Discuss a project