AI for Ocean Climate Modelling: 12-month Industry Placement

Met Office; Devon and Torbay; Hybrid; GBP 24,000 - 30,000; Full time

Benefits offered by this job

27.5 days annual leave

Discounted shopping

Salary £27,170

Job summary

Met Office is offering an Industrial Placement for 2027 in AI for Ocean Climate Modelling. The 12-month Year in Industry placement starts July 2027, with potential hybrid working where office attendance is encouraged at least once a week.

You will join the Ocean Modelling team within the OCC group at the Met Office Hadley Centre, contributing to data-driven approaches for ocean vertical mixing and gaining skills in machine learning to improve model accuracy.

Qualifications

• Lead criteria: Experience with programming or scientific computing tools (Fortran, Python, MATLAB).
• Degree in mathematical/physical science with industrial placement year included.
• Strong interest in AI and machine learning and ability to learn quickly.
• Strong analytical and problem-solving skills.
• Clear written and verbal communication of technical ideas.

Responsibilities

• Review literature on ML approaches to ocean vertical mixing.
• Select and process training datasets for ML.
• Develop and evaluate a data-driven approach for ocean vertical mixing.
• Assess performance using independent data or simple numerical experiments.
• Analyze, document and communicate results, including limitations and recommendations.

Skills

Fortran

Python

MATLAB

AI/ML interest

Education

Degree in a mathematical or physical science with industrial placement year

Job description

Met Office is offering an Industrial Placement for 2027 in AI for Ocean Climate Modelling. The 12-month Year in Industry placement starts July 2027, with potential hybrid working where office attendance is encouraged at least once a week.

You will join the Ocean Modelling team within the OCC group at the Met Office Hadley Centre, contributing to data-driven approaches for ocean vertical mixing and gaining skills in machine learning to improve model accuracy.