Geoscience Data Assistant

About the role

Join IMDEX as a Geoscience Data Assistant and play a key role in supporting cutting-edge geoscience and machine learning solutions. You'll apply your geoscience expertise to create and manage high-quality datasets, validate geological imagery, and support projects that drive innovation across geology, geotechnics, computer vision, and AI. 

Key responsibilities

• Accurately annotate imagery using third-party or in-house annotation software to create high-quality training and evaluation datasets

• Manage standard and advanced Tech Evaluations using Datarock general and unique models and prepare standardised and bespoke summary outputs

• Assist with the sampling, processing and delivery of general model and bespoke customer products and inference outputs

• Prepare, correct, and organise standard and more complex imagery and drill hole metadata for use on the Digital OBK platform

• Undertake validation of annotated datasets, ensuring accuracy, completeness, and adherence to project guidelines

• Identify and correct errors or inconsistencies in annotations

• Collaborate with Data Coordinator, Geologists and Geotechnical Engineers, Data Scientists, ML Engineers, and project teams to understand data requirements.

About you

• Certification in Geology, Geotechnical Engineering, or a closely related discipline

• Exposure to geoscience, mining, drilling, or geological imagery

• High attention to detail and accuracy in technical task execution

• Proficiency with computers and image editing tools

• Experience with Microsoft Office Suite or similar, especially spreadsheets

• Basic familiarity with computer vision concepts or terminology

• Ability to manage multiple tasks and work to structured deadlines

• Strong written and verbal communication skills, including the ability to create and maintain training documentation

• Proven ability to work both independently and collaboratively

• Desirable: Experience using computer vision annotation tools (e.g., V7, CVAT, Labelbox, Supervisely)