About the role
AirCredit in Berlin is seeking a Machine Learning Engineer (Document AI) to own models and pipelines that read, classify and translate financial documents into structured data for a bank-ready output.
You will collaborate with engineering and product teams, define quality metrics and take your models to production, with a strong emphasis on privacy and EU compliance. Fluency in German and English is expected.
Build and improve models for document classification and data extraction (OCR, layout, NLP). Apply and fine-tune LLMs for extraction, plausibility checks and structured outputs. Build robust data and evaluation pipelines on real, messy documents. Own accuracy, latency and reliability in production. Care for privacy and security — we handle highly sensitive financial data in the EU. More than 3 years of experience with machine learning in production systems. Strong Python skills and experience with modern ML frameworks (e.g. PyTorch). Hands-on experience with NLP, LLMs or document/computer-vision processing. A feel for solid evaluation over gut feeling — you measure before you claim. Good German and confident English. Experience with OCR, layout models, RAG, MLOps or regulated industries like FinTech. Not required — curiosity and rigour count more. Equity 30 days of holiday