🚀 Be part of a movement to change the way Europe pays

In today’s digital world, payments often still feel outdated: random delays and confusing rules make it harder than it should be to pay and get paid. The European Payments Initiative (EPI) is here to change all that, forever.

With Wero, our digital wallet, we make sending and receiving money simple, seamless and secure across France, Belgium and Germany, with more countries and omnichannel solutions coming soon. Supported by 14 major banks and the two largest European acquirers, EPI is building a new, proudly European payment system: easy, instant and transparent, all for the greater good.

🔎 What's in it for you

We are currently on the lookout for an NLP Data Engineer with hands-on experience in text data processing and document ranking. You’ll join a team working on heterogeneous data sources, where your work will help establish processing pipelines that optimize document relevance and retrieval. 

🐝 About the team

You’ll join a pan-European AI team directly reporting to the CTPO of the company. The team is dedicated to building reliable and impactful GenAI solutions and operates in an Agile environment and works closely across countries to design, benchmark, and deploy LLM and RAG systems that truly make a difference. They are located across Europe, in Germany, France, the Netherlands and Belgium.

💥 Your impact

• Prepare, clean, and sanitize heterogeneous text data sources (Confluence, Readme, Slack, GitHub) for usage in AI applications 
• Design and implement document scoring and ranking pipelines using models like BM25, TF-IDF, or neural rankers 
• Build and optimize embedding workflows to support semantic similarity, nearest-neighbour search, and RAG pipelines 
• Research and benchmark the latest NLP methods to continuously improve document processing, retrieval, and ranking

💻 Technology stack

• Languages & frameworks: Python, PySpark, Numpy, Pandas
  
• NLP libraries: spaCy, NLTK, Hugging Face Transformers
  
• Embedding models: Word2Vec, GloVe, FastText, BERT, RoBERTa, LASER
  
• RAG frameworks: LangChain, LlamaIndex, Haystack, Guidance, PromptLayer
• Infrastructure & deployment: AWS, Docker, Kubernetes

🕵🏻‍♀️ To succeed, you should meet at least 70% of these requirements

• +3 years’ experience in building NLP and text processing pipelines. An advantage would be prior experience in AI and ML in general. 
• Ability to understand, design, and apply evaluation metrics for document ranking and retrieval  
• Experience with embeddings and similarity search, including pre-trained and fine-tuned models such as BERT, Word2Vec, or FastText 
• Experience implementing document scoring and ranking models (BM25, TF-IDF, or neural rankers)  
• Familiarity with RAG frameworks such as LangChain, LlamaIndex, or Haystack
• Proficiency in Python and NLP libraries like spaCy, NLTK, Hugging Face Transformers  
• Ability to evaluate, design, and implement improvements to document processing and retrieval workflows
• Fluent in English (CEFR C1 or C2);French, German, Dutch, or other European languages are a plus.

🪜 If this looks like you, the recruitment steps are:

1. A first call with one of our recruiters
2. An interview focused on the mission and your expertise with your future manager
3. A final interview with our CTPO
4. Hopefully, an offer you can’t refuse

⛔ Turn back if …

• You’re not comfortable getting hands-on with messy, heterogeneous text data and turning it into structured, usable inputs
  
• You lack a solid understanding of document ranking, embeddings, and the importance of numbers-driven evaluation
• You're not confident using both pre-built NLP tools and custom created NLP and text processing pipelines

Otherwise apply!

🫶 Our commitment to equal employment opportunities

EPI offers the same job opportunities to all, without distinction of gender, ethnicity, religion, sexual orientation, social status, disability or age. EPI promotes the development of an inclusive work environment that mirrors the diversity of the clients our product is serving.