ChristophC.Stach

$ cat curriculum.md

Curriculum

A concise overview of my professional experience and education: over a decade of full-stack engineering with Vue/Nuxt, React/Next.js, and NestJS, more recently extending into Python for AI and LLM agents at MBition (Mercedes-Benz), and a B.Sc. and M.Sc. in Applied Computer Science focused on AI, bioinformatics, deep learning, and data science.

Download CV (PDF)

experience

  1. Senior Full Stack Engineer @ MBition GmbH (Mercedes-Benz)

    Berlin

    • TypeScript
    • Vue
    • Nuxt
    • NestJS
    • Python
    • FastAPI
    • LLMs
    • Model Context Protocol (MCP)
    • Kubernetes
    • PostgreSQL
    • MongoDB
    • Built the fully typed Vue datatable behind every list screen, with server-side pagination, sorting, and configurable filters that compile into a single database query, replacing tables that had been hand-built per screen
    • Built SDP, a no-code platform for configuring and deploying Mercedes-Benz speech applications, and ship features on it end-to-end in Vue and Nuxt with strict TypeScript, zod-validated REST contracts, and realtime updates over WebSockets
    • Built a Role-Based Access Control (RBAC) system with CASL.js, enforcing one permission definition across both the Nuxt UI and the NestJS API
    • Consolidated multiple products into a pnpm monorepo and hardened type safety across the TypeScript codebase, sharing types end to end and enforcing them in CI, with build times down ~50%
    • Build and operate the Python platform that hosts our LLM agents on Kubernetes, providing the shared FastAPI service, base classes, and CI/CD that agent teams build on
    • Drove AI-assisted development across the team: authored the AGENTS.md instruction set and reusable agent skills, and set up a team wiki in the Open Knowledge Format (OKF) to make organizational knowledge agent-consumable
    • Re-architected the platform from synchronous to fully asynchronous Python, cutting memory consumption by 75% (4 GB to 1 GB) and removing the need for multiple worker processes
    • Decoupled the codebase with dependency injection so components can be mocked and unit-tested in isolation, then enforced quality gates in CI (Ruff linting, ty static type checking, and a pytest suite), surfacing and fixing ~5,000 pre-existing type errors
    • Migrated the project from GitHub to GitLab and parallelized the CI/CD pipelines, halving runtime; each merge request now spins up its own Kubernetes pod, giving 140 contributors a production-like environment to validate changes in

  2. Frontend Engineer @ DERICON GmbH

    Berlin & Frankfurt

    • Vue
    • Nuxt
    • React
    • Remix
    • TypeScript
    • Angular
    • Launched a graph-based no-code workflow builder (React/Remix) that let non-technical users model and deploy workflows independently
    • Delivered new features on an Angular-based frontend, modernizing components and improving performance and accessibility
    • Built middleware and UI integrations to bridge legacy systems with modern REST APIs, simplifying data flows and de-risking migrations

  3. Full Stack Developer @ SLH GmbH

    Schüttorf

    • JavaScript
    • ExtJS
    • REST
    • PHP
    • Symfony
    • MySQL
    • Modernized a legacy frameset-based PHP application into a modular front-end/back-end architecture, rebuilding the UI with ExtJS on a clean PHP backend
    • Collaborated with stakeholders to prioritize requirements, deliver new features, and resolve defects quickly

education

  1. M.Sc. Applied Computer Science @ HTW Berlin

    • Thesis: Explorative Analysis of Data from Nanopore-Based DNA Sequencing to Identify Dilution Effects
    • Specialisations: AI, Deep Learning, Data Science, Bioinformatics
    • Grade: 1.1
    • Exchange: NTUST, Taipei

    Read the thesis (PDF)

  2. B.Sc. Applied Computer Science @ HTW Berlin

    • Thesis: Style transfer for photos using deep learning
    • Grade: 1.6
    • Exchange: UNINTER, Cuernavaca, Mexico

  3. IT Management Assistant (Apprenticeship) @ Kortmann Beton GmbH & Co. KG

    • Completed apprenticeship alongside theoretical studies
    • Grade: 2.0

projects --recent

  • › TIGeR (Toolkit for Integration of Generative AI Resources)

    The common runtime multiple Mercedes-Benz agent teams build their LLM agents on, with Agent-to-Agent and Model Context Protocol interoperability.

    • Python
    • FastAPI
    • LLMs
    • A2A
    • MCP
    • Kubernetes
    • GitLab CI

  • › SDP (Speech Development Platform)

    No-code platform for configuring and deploying speech applications for Mercedes-Benz.

    • Vue
    • Nuxt
    • Tailwind
    • NestJS
    • PostgreSQL
    • MongoDB
    • Websockets

  • › Crowds UI

    Workflow editor with graph-based task views and dynamic form generation to streamline internal processes.

    • React
    • Remix
    • Tailwind