Model Risk Data Scientist
Posting date: | 29 July 2025 |
---|---|
Hours: | Full time |
Closing date: | 28 August 2025 |
Location: | Edinburgh, EH12 1HQ |
Company: | NatWest Group |
Job type: | Permanent |
Job reference: | R-00261756-OTHLOC-GBR-5FEDI034 |
Summary
Join us as a Model Risk Data Scientist
- If you have experience building and validating machine learning models, this is a fantastic opportunity join our innovative, vibrant team in our Risk function
- You’ll be performing technical reviews and oversight of models used in the bank, and working with model development teams to continually drive up the value generated by data-driven modelling
What you'll do
This Model Risk Data Scientist role will see you reviewing and independently validating assigned models in accordance with the bank’s policies and model standards. You’ll communicate your findings and recommendations to stakeholders and advise on how model risk can be reduced or mitigated.
As well as this, you’ll be developing solutions for automating validation activities while understanding model and data usage, quality and interdependencies across the bank.
Your role will also involve:
- Developing the team's analytics codebase, adding functionality, fixing issues and testing code
- Support the development of a validation framework for Gen AI models across the bank
- Reviewing your colleagues’ analysis, code and reports
The skills you'll need
We’re looking for someone with an excellent grasp of mathematical methods, concepts and assumptions that underpin machine learning, statistical modelling and artificial intelligence.
You’ll also need:
- A proficiency in Python and libraries commonly used for data science, such as Linux WSL and AWS Sagemaker
- Practical experience building and validating AI models particularly LLMs
- The ability to extract the essential ideas underlying technical results and explain them in terms of their practical consequences
- Excellent technical and communication skills
- The ability to deal with ambiguity and to work autonomously