We’re seeking for a self-driven, proactive and startap-minded AI Engineer to join a German healthcare project — an AI medical “brain” for doctors and medical professionals that makes billing faster, more accurate, and fully compliant. The platform leverages state-of-the-art language models to interpret clinical notes, recommend CPT, ICD-10, and HCPCS codes, and check compliance against payer and regulatory rules.
What You’ll Do * Build and fine-tune large language models (LLMs) for medical text understanding * Implement retrieval-augmented generation (RAG) pipelines with vector databases * Train and evaluate models on real clinical notes, CPT, ICD-10, and HCPCS coding data * Incorporate logic for payer-specific and regulatory coding rules (bundling, modifiers, and regional/state variations) * Optimize inference for speed and accuracy (quantization, distillation, GPU tuning) * Collaborate with product + medical experts to turn research into production features * Establish scalable workflows and documentation so the team can quickly adapt to updated LLMs, datasets, and payer/regulatory requirements
What We’re Looking For * Strong experience in NLP/LLMs (PyTorch, Hugging Face, or equivalent) * Background in retrieval systems, vector databases, and RAG pipelines * Proven ability to train and fine-tune models on specialized datasets * Ability to design processes for precision, accuracy, and rigorous benchmarking to ensure models meet compliance and real-world medical billing standards.
Nice to have * Understanding of healthcare coding standards (CPT, ICD-10, HCPCS) * Must have past projects we can review (GitHub, online demos, or similar) * Bonus: experience with model optimization for local inference (quantization, GGUF, ONNX, etc.)
Why Join * Tackle Hard Problems: Work on cutting-edge challenges at the intersection of AI, healthcare, and compliance — where every improvement directly reduces denied claims and improves patient care. * Shape the Product: Your expertise won’t be a small cog in the machine — you’ll own critical decisions in model design, fine-tuning, and RAG architecture. * High-Velocity Startup: Move fast, experiment boldly, and see your work ship to production in weeks, not quarters. * Mission-Driven Team: Join people who care deeply about fixing one of the most painful problems in U.S. healthcare with technology that’s practical, ethical, and transformative.
