Stefan Baum

Senior Software Engineer · System Architect · DevOps


Profile

20+ years of experience in full-stack development, system architecture, and DevOps. All-rounder: distributed systems, legacy modernization, containerization.

Since 2025 as private R&D: local AI infrastructure for open-weight LLM inference and AI-integrated workflows.

Industries so far: enterprise IT, automotive, public sector, e-commerce, travel.


Experience

Since 2019 — Freelance Consultant & Senior Software Developer

  • Embedded Platform — In-car infotainment for premium automotive; PM, prototyping, multi-gen support (2019–2024)
  • E-Government Digitalization — Baden-Württemberg, Angular 14, NgRx (2022–2023)
  • Booking Platform Modernization — Migration of legacy Java services from vendor VM hosting to a containerized Docker architecture; discovery, staging, production cutover, and post-migration stabilization (2025–2026)
  • CMS Migration (2025–2026)
  • Containerized Environment Setup — Reproducible dev, staging, and production environments for the platform's services: Docker Compose, Traefik routing, Elasticsearch, CI/CD pipelines, and deployment runbooks (2025–2026)
  • Dev-Team Consulting — Knowledge transfer, technical documentation, and ongoing support for the client's development team (2026)
  • E-Commerce Data Integration — B2B data aggregation, MuleSoft (2020)
  • Local AI Platform — Private R&D: open-weight LLM inference on NVIDIA Ampere-class GPUs, a dual DGX Spark GB10 pair in tensor-parallel mode, and Apple Silicon; engines vLLM, llama.cpp, MLX; one OpenAI-compatible LiteLLM gateway that routes by privacy class; RAG on PostgreSQL/pgvector; MCP integrations; models and inference engines kept current as they appear (2025–2026)
  • Open-Weight Model Evaluation — Private R&D: comparison of current open-weight models for development and document-heavy workloads: quantization (NVFP4, GGUF), context length, throughput, and tool-calling reliability; containerized serving monitored with Prometheus and Grafana (2025–2026)
  • Document Intelligence Pipeline — Private R&D: a staged pipeline that ingests business documents, extracts content with OCR and vision models, classifies them, and files them by taxonomy into a DMS and a RAGFlow index; LangGraph workflow engine (2026)

2017–2019 — Senior Software Developer, Publikat (Signa Sports United)

  • B2B data aggregation and e-commerce platform development
  • MuleSoft integration and Navision data provider development
  • System integrations and ETL pipelines

2015–2017 — Senior Software Developer, Lifestyle-Travel-Team GmbH

  • Java-based reservation system for bike tourism
  • Full-stack: booking platform, admin, website, tech support

2005–2015 — Software Developer, maxviva Technologies AG

  • Full-stack internet booking engine with GDS backend integration
  • High-volume transaction platform with 24/7 operations
  • Performance optimization and high-availability systems
  • Java / Spring

Technologies

AreaStack
BackendJava, Spring Boot, FastAPI, Node.js, Python
FrontendAngular, TypeScript, NgRx, RxJS
DevOpsDocker, CI/CD Pipelines, Linux, Infrastructure as Code, GPU Server Operations (private)
DatabasesPostgreSQL, MySQL, Elasticsearch, Redis, Vector DBs
AI/ML (private)LLM Integration, Local Inference, vLLM, llama.cpp, MLX, LiteLLM, RAG Pipelines, pgvector, Model Evaluation, Document Intelligence, LangGraph, MCP
ToolingClaude Code, OMP, OpenCode, Pi, Hermes, VS Code, Cursor, Playwright, LangChain, Context Engineering, Coding-agent evaluation
ObservabilityGrafana, Prometheus, Logging Stacks
ArchitectureSystem Integration, Event-Driven, API Design

Education

Diplom-Informatiker (FH) — Fulda University of Applied Sciences, Applied Computer Science (1998–2004)


Languages

German (native), English