About the role
Senior Python Engineer - Email Analytics & Taxonomy Pipelines | Luxembourg
📍 Luxembourg (on-site) | 📅 ASAP | ⏳ 3-month project | 💼 Freelance
At PIXIE Luxembourg, we’re looking for a Senior Python Engineer to build the historical email analysis pipeline for an important project with an insurance client in Luxembourg.
The client already knows the main reasons customers get in touch. What it needs is a way to analyse historical emails and find out what each type of request actually contains. Which documents usually come with it ? What distinguishes it from a similar request ? Which fields should be extracted, and how should they be checked ?
Your pipeline will turn those findings into an editable reference that business teams can review and correct and that the document analysis engine can use directly. Adding a new request category should mean rerunning the pipeline or editing the reference, not writing new Python code.
This is an intensive three-month project. Your pipeline needs to work on an initial set of categories and feed the first usable package by the end of week three. You’ll then expand it across the full taxonomy using the same approach.
🎯 Your mission
- Analyse historical emails and attachments against the client’s existing taxonomy. Identify recurring document types even when the original attachment labels are too broad to be useful.
- Build a reusable pipeline that generates category and document descriptions, characteristic terms and spelling variants, expected attachments and distinctions between closely related requests.
- Define what the extraction engine needs. Generate a readable JSON or YAML reference stating which fields to extract, where to look for them, how to normalise them and which consistency checks to apply.
- Make the reference usable by both engineers and business teams. Agree the schema with the Document AI lead, validate the generated output, and make proposed rules easy for business users to review and correct.
- Take it from a subset to full coverage. Test on real historical emails in the client’s environment, incorporate business feedback and expand across the full taxonomy without writing category-specific logic.
🔎 Who are we looking for?
- You’ve built maintainable Python pipelines that turn unstructured text or documents into usable, structured outputs.
- You’ve worked on text classification, taxonomy-driven processing or document analytics, and can explain how you checked that the classifications or generated rules made sense.
- You’re comfortable designing and validating JSON or YAML schemas and keeping business rules separate from application code.
- You can investigate messy historical data, deliver a working first version quickly and work on-site in Luxembourg.
Experience with email or attachment parsing, NLP or AI-assisted classification is a plus. The key is being able to show how your pipeline handles new categories without a rewrite.
🤝 Interested ? Let’s connect !
👉 Apply directly through this LinkedIn job posting with your updated CV, earliest availability and expected daily rate. In a few lines, tell me what pipeline you built from unstructured data and how you handled new categories or changing business rules.
Questions? Feel free to message me directly.
#Freelance #Luxembourg #Python #DataEngineering #NLP #PIXIE