Full-Stack AI Engineer — I build production AI platforms for higher education, and the infrastructure that serves them.
Currently at the University of Wuppertal (ZIM), where I develop and operate the KI-Agenten-Studio — an AI assistant platform for teaching and research: 530+ REST endpoints, unified abstraction over 6 LLM providers, hybrid RAG, multi-agent orchestration, and MCP integration.
🔬 Featured: llm-inference-lab
A university-scale LLM serving platform in miniature — built, measured, and documented end to end:
| Layer | What's inside |
|---|---|
| Serving | vLLM (Qwen3-8B) on a cloud GPU — concurrency sweeps, TTFT/ITL percentiles, saturation analysis |
| Governance | LiteLLM gateway: virtual keys, per-faculty budgets & rate limits, chargeback pricing, a live-proven model fallback chain |
| Operations | Prometheus/Grafana as code, SLOs derived from my own benchmarks, alert rules with a demonstrated firing case, runbook |
| Platform | Kubernetes (k3d) + Argo CD — the whole stack deployed and scaled via GitOps |
Every number is measured, every operational failure is documented in the journal.
AI / LLM 
Languages 
Backend / Frontend 
Platform / Ops 
XR 
- KI-Agenten-Studio — AI assistant platform for teaching & research (University of Wuppertal)
- Prompt Coach — prompt competence platform, 230+ prompt-engineering techniques, WCAG 2.1 AA (university edition)
- Innovation Board — schema-driven platform for XR/maker projects with a self-hosted .NET AI subsystem (second instance)
- tcrijschool.nl · speedwaycar.nl — freelance full-stack builds
- "KI-Assistenten-Studio: Building single- and multi-agent systems" — 12-session online training series for university teaching and research staff (2026)
- "From AI assistants to AI agents — the next generation of intelligent systems" — university-wide continuing-education session (2026)
📫 firas.alaammar@gmail.com · Essen, Germany