Large?'scale data migrations at NN must not only transform and clean massive datasets but also prove, with full auditability, that the numbers still add up. Manual reconciliation is slow, error?'prone, and costly, especially under the scrutiny of DNB audits. This thesis tackles that challenge by designing an AI?'supported framework that automatically verifies numeric and provenance consistency across medallion?'layer migrations (bronze ?' silver ?' gold). The goal is to detect and explain aggregation mismatches, infer transformation impacts, and produce cryptographically verifiable evidence packages, from lineage graphs to signed aggregate proofs, that are both human?'readable and audit?'ready. By combining statistical analysis, machine learning, and provenance tracking, the resulting prototype aims to deliver faster, reproducible, and trustworthy audit artifacts, dramatically reducing manual effort while strengthening NN's compliance posture.
What you are going to do
You will develop and validate an AI?'powered reconciliation framework that ensures numeric and provenance consistency across NN's medallion?'layer data migrations, delivering audit?'ready proof for compliance with DNB requirements. The solution may combine statistical analysis, machine learning, provenance tracking and, where beneficial, autonomous agents to achieve its goals.
-
Design methods to detect and localize aggregation mismatches after filtering, cleaning, and transformations
-
Quantify and explain expected vs. Unexpected data loss in migration pipelines
-
Integrate statistical, ML, and provenance?'tracking techniques into a modular prototype, optionally enhanced by agents
-
Generate human?'readable and cryptographically verifiable audit artifacts (lineage graphs, signed proofs, anomaly explanations)
-
Evaluate the framework on historical NN migrations and synthetic benchmarks to measure accuracy and audit?'effort reduction
Who you are
You have a strong passion for solving complex analytical challenges and applying AI technologies to improve processes in the financial and risk domains. You work in a structured way, think critically, and pay close attention to regulatory requirements and technical specifications. In addition, you bring:
-
A degree or specialization in Computer Science, Data Science, Artificial Intelligence, Econometrics, or a related field
-
Hands?'on experience with Python
-
Solid knowledge of statistics, machine learning, natural language processing, and AI agent frameworks
-
The ability to translate complex technical concepts into clear, well?'structured documentation and actionable solutions
-
Self?'motivation and independence in planning and conducting research
Who you will work with
You will be part of the Data & AI team within the CIO department, a growing group of specialists who share a passion for creating impactful data and AI solutions. Our team offers plenty of room for learning and personal development. In this project, you will collaborate closely with colleagues from IT, business, and support departments to design and deliver AI-driven solutions for regulatory document generation.
Your employer will be NN Group, an organization that offers a wide range of opportunities. You will work within the Life & Pensions business unit, focusing on innovation in compliance and risk management processes.
Any questions?
You can reach out to Rowena Taal, Internship coordinator, via stagebureau@nn.nl.
Are you unsure if you meet 100% of the requirements? We encourage you to apply anyway and show us the unique value you could bring!
Het salaris bedraagt €650