AI Engineer – Python / LLM & Extraction documentaire

Your day-to-day

  • Exploring a large corpus of incoming emails to automatically identify the different families of requests (clustering, topic modeling, embeddings, LLMs)
  • Structuring a reference framework of request types and of the data expected for each family, in close collaboration with the internal AI team
  • Designing and training classification models to route incoming emails according to this framework
  • Extracting data reliably from email bodies and attachments (PDF, scanned documents, forms) using OCR, document parsing and LLM-based structured extraction
  • Developing the end-to-end pipeline in Python and exposing the services through APIs (FastAPI)
  • Defining quality metrics (precision, recall, extraction accuracy) and iterating to make the results reliable
  • Working on-premise with confidential data that cannot leave the client’s environment, then documenting the solution and ensuring the handover to the internal team

Your profile

  • At least 5 years of experience in Python on Data Science / AI projects, including FastAPI or similar frameworks
  • Proven experience in OCR and document extraction (Tesseract, PaddleOCR, docTR, Azure Document Intelligence or equivalent)
  • Strong NLP background: text classification, clustering, embeddings, Transformers (Hugging Face)
  • Hands-on experience with LLMs / Generative AI: structured information extraction, prompt engineering, RAG
  • Experience deploying open-source LLMs on-premise (vLLM, Ollama, Mistral, Llama) is a strong asset; insurance or banking experience is a plus
  • Fluent in French and English, both written and verbal; German or Luxembourgish is a plus
  • Autonomous, pragmatic and solution-oriented, with the ability to structure an exploratory topic
Job Category: Artificial Intelligence / Data Engineering
Job Type: Consulting
Job Location: Luxembourg

Apply for this position

Allowed Type(s): .pdf, .doc, .docx