About this role
## Deine Aufgaben:
We're looking for a hands-on AI engineer who adapts open-source models so our platforms, for example our tax advisory group, can develop and operate their own specialised models.
- Select, fine-tune and evaluate open-source LLMs (e.g. Llama, Mistral, Qwen) for domain-specific use cases such as tax, accounting and advisory workflows
- Build reusable pipelines for data preparation, fine-tuning (LoRA/QLoRA), RAG and model evaluation that can be rolled out across multiple platforms
- Deploy and run models on sovereign, GDPR-compliant infrastructure, with confidentiality requirements of regulated professions built in from day one
- Work directly with platform management and domain experts to turn real workflows into working AI products
- Enable platform teams to maintain and further develop their models independently
## Was Du mitbringst:
• Several years of experience training, fine-tuning or deploying LLMs in production • Strong Python and hands-on experience with PyTorch, Hugging Face, vLLM or similar • Solid understanding of MLOps, GPU infrastructure and model evaluation • A builder mindset: you enjoy moving from prototype to production quickly • Fluent English; German is a strong plus
## Was wir bieten:
• Join our Core AI Team and shape the AI strategy of an entire portfolio from the start • Real impact: your models go live in companies with thousands of employees and clients • Entrepreneurial environment with short decision paths and direct access to leadership • Munich-based, with flexible working