Functie
Data Scientist GenAI / Chatbot
Location: Utrecht
Level: Medior–Senior
Working arrangement: Hybrid, with Thursdays in the office at Winthontlaan, Utrecht.
About the role
As a Data Scientist (GenAI / Chatbot), you’ll help improve the quality and effectiveness of chatbot systems. You’ll work across data, engineering and product, investigating how chatbots behave, identifying weaknesses and turning your findings into meaningful improvements.
Your focus is on understanding why a chatbot isn’t behaving as expected and what context or structure it needs. You’ll assess current solutions against state-of-the-art approaches and contribute to innovation in retrieval, generation, reasoning loops, tool calling and related GenAI capabilities.
What you’ll do
Analyse chatbot outputs and identify improvements with the greatest impact.
Diagnose context gaps and determine what additional information agents need.
Work with product managers and engineers to define criteria for evaluating agent performance.
Partner with use-case teams and stakeholders to understand real-world scenarios and turn feedback into improvements.
Improve retrieval mechanisms, data structuring and the onboarding of contextual sources.
Design evaluation frameworks that support rapid, large-scale experimentation and continuous innovation.
Translate functional requirements into actionable technical improvements.
What you bring
A strong analytical mindset and a passion for understanding AI system behaviour.
A strong understanding of GenAI systems and experience with LLM-based applications.
Experience with retrieval-augmented systems.
A strong understanding of prompts, context windows, retrieval and grounding.
The ability to evaluate output quality and explain why results are effective or ineffective.
Familiarity with API integrations, batch synchronisation, and structured and unstructured data.
Comfort collaborating closely with engineers and troubleshooting data flows and context ingestion.
The ability to design evaluation frameworks beyond traditional machine learning metrics.
An understanding of functional context and user behaviour, and the ability to work across engineering and product.
Clear communication skills: you can explain context gaps and reasoning to engineers, product managers and stakeholders, and create effective improvement proposals.
Nice to have
A background in human-in-the-loop evaluation.
Who will thrive in this role
You enjoy figuring out why AI systems behave the way they do and how they can be improved. You like debugging behaviour as well as code, think structurally, connect the dots and have a strong intuition for what good performance looks like.
Eisen
- The candidate must be able to work in English
- Midlance constructions are not allowed
- The candidate must be available throughout the entire duration of the assignment
- The candidate must work in the office on Thursdays at Winthontlaan, Utrecht, until further notice
- The candidate must submit their CV and personal motivation in English
- The candidate must be submitted exclusively to Magnit – Rabobank during the four-business-day exclusivity period for one request
- The candidate must possess a valid passport or ID card and bring it to the interview and at the start of the assignment
- The candidate may start only after successfully completing the pre-employment screening
Gunningscriteria
- Background in human-in-the-loop evaluation is a plus
Extra informatie
Je sollicitatie dient uiterlijk 28 oktober 2026 om 17:00 uur door ons te zijn ingediend, voorzien van een motivatie (waarin wordt toegelicht hoe aan de eisen en gunningscriteria wordt voldaan) en twee referenties. Wij ontvangen je gegevens graag minimaal één werkdag vóór de sluitingsdatum.
