Matthias Vongerichten

Freelance Software & Data Engineer, Berlin, DE

tl;dr: LinkedIn || Malt

About Me

I built Mercedes-Benz 🚗 software for almost a decade, then spent nearly four years integrating data 📄 at ARMEDANGELS 👼. Since November 2024 I work for myself.

Most of what I do now is AI agents 🤖 and automation. I run Von Court Agents, a fleet of specialist agents running on EU-hosted models 🇪🇺 with no US cloud in the loop — and I am currently building an emissions accounting platform for Zukunftswerk eG, end to end. I like Python, TypeScript, PostgreSQL and Kubernetes.

I choose projects by the sustainability impact they can have 🌿 — that is my filter, not my niche. Whatever the domain, the work still has to be engineered properly.

Want me for freelance work? You can contact 📧 me in english, german and you can also try hungarian.

"We must all suffer one of two things: the pain of discipline or the pain of regret or disappointment."

– Jim Rohn

Services

💪 I'm good at...

  • AI Agents & Automation
    • Agent fleets and orchestration (Claude, MCP, n8n)
    • LLM integration – retrieval, embeddings, evaluation (Python)
    • Workflow Automation (Airflow, Python, n8n)
  • Data Integration
    • Data Processing & Replication (Airflow, Python)
    • Data Modeling (dbt, Snowflake, Looker)
  • Backend Development (Python, FastAPI, Java, PostgreSQL)
  • Infrastructure Automation & Self-Hosting (Kubernetes, Docker)
  • Research & Proof of Concepts
  • Technical Product Ownership/Management (Jira, Confluence)

🧪 I'm experimenting with...

  • Sovereign AI – EU-hosted and local models as the default, not the fallback
  • Long-lived memory for agents across sessions
  • ... and a lot more!

Experience

Freelance Software Engineer – - Present

  • Built an emissions accounting platform end to end – sole engineer
  • Web application and UI (Next.js, React, TypeScript)
  • Backend services and APIs (FastAPI, Python)
  • Data model and processing pipelines (PostgreSQL, Python)
  • Embedding-based matching of activity data to emission factors (Python)
  • Containerised deployment across development, staging and production (Docker)
Next.js TypeScript FastAPI Python PostgreSQL Docker
Placeholder image

ARMEDANGELS

Data Engineer – -

  • Introduced unit tests into the Airflow environment (Python)
  • Refactored Airflow DAGs and Operators (Python)
  • Transformed XML files (XSLT, Python)
  • Created and maintained a system landscape (diagrams.net, UML)
  • Built BI reports and dashboards (Looker, Excel, DAX)
  • Maintained and extended the data model/data warehouse (dbt, Snowflake, SQL)
  • Integrated several partner services (Python, Airflow, WSDL, REST)
Airflow Python UML Looker dbt Snowflake
Placeholder image

Mercedes-Benz

Software Engineer & Technical Product Owner - -

  • Technical Product Owner of the MB Vans Configurator for potential end customers – owned the product and its technical direction (Java, TypeScript)
  • Technical Product Owner Proxy of a vehicle configuration and offer creation application, critical to MB Vans Sales
  • Oversaw the modernization of the MB Vans Configurator tech stack, directing the developers who built it:
    • Migration from Oracle to PostgreSQL DB (Java, Spring Boot)
    • Migration from a legacy application for data manipulation to a low-code platform (Retool, JavaScript, SQL)
    • Automation of data processing and software deployments (Jenkins, Java, Shell)
    • Migration to a container orchestration platform (Kubernetes)
  • Managed the technical side of adapting and operating a modified version of the MB Vans Configurator API for an internal customer (Java, Spring Boot, Git)
  • Managing software development in Scrum-style, and coordinating user testing (Jira, Confluence)
  • Planning and execution of the rollout of software versions in various countries (Jira, Confluence, GitHub)
  • Ascertaining of customer requirements, and creation of epics and stories (Jira)
  • Giving presentations on future software features to international audiences (PowerPoint)
  • Driving the rollout of usage analytics across multiple applications (Adobe & Google Analytics)
  • Acting within the department as an expert for a complex vehicle configuration interface (Java, Postman)
  • Managing external development team
  • Creation and ongoing maintenance of comprehensive representation of the department-wide system and service landscape (diagrams.net)
  • Developed the scenarios for modernizing legacy applications (micro frontends, microservices)
  • Drove the introduction of E2E monitoring for the production environment (Jenkins, Protractor, TypeScript)
  • Designing an interface for an internal financing service (OpenAPI)
  • Introduced a smoke test automation process built on SikuliX (open source, Python, OCR-only)
  • Administration of Jira, Confluence and SharePoint instances
Java TypeScript Spring Boot PostgreSQL Kubernetes Jenkins Retool
...

Projects

Von Court Agents

A fleet of specialist AI agents, self-hosted

A set of agents that each own one job – infrastructure maintenance, bug fixing, email, UI work, coordination – instead of one assistant that does everything badly. They share a memory layer, a credential broker that isolates every agent's secrets from the others, and a task board they pull work from.

Two constraints shape the design. Everything runs on EU-hosted models with no US cloud in the loop, so the data stays under European jurisdiction. And model choice weighs emissions per task as a first-class criterion, not an afterthought – the smallest model that can do the job gets it.

Python MCP Docker Matrix Notion API

Education / Training

Master of Computer Science (on hold)

IU International University of Applied Sciences, Germany; part-time; approx. 40% completed
-

Bachelor of Science – Business Informatics

AKAD University Stuttgart, Germany; part-time; grade 2.1
-

Courses

Data Analyst Nanodegree

Udacity

Professional Scrum Master I

Scrum.org

Contact