AI Data Trainer – Multi-Modal LLMs (On-site, Den Haag)

Jobster; Den Haag; On-site; EUR 37,000 - 46,000; Full time

Benefits offered by this job

Fixed-term contract

Office-based, Monday to Friday

Regular trainings and events

Contribute to advanced AI projects

Experience with LLMs and data analysis

Job summary

Amazon in The Hague is seeking an AI Data Trainer to create, annotate, and review training data across text, image and video. You will write texts in multiple styles following guidelines and perform quality checks to improve language models.

You will collaborate with scientists and engineers to refine guidelines, analyze data, identify issues, and propose solutions. The role focuses on data and tool improvements while handling sensitive content.

Qualifications

• Dutch and English proficiency (C1)
• Strong analytical skills
• Attention to detail
• Experience with annotation and editing language data
• Understanding of high-school STEM topics
• Ability to adapt in a fast-paced environment

Responsibilities

• Create, annotate, and review high-quality training data across text, image and video
• Write texts in multiple styles according to guidelines
• Perform quality checks on data and tooling
• Collaborate with scientists and engineers to improve guidelines
• Improve performance of Large Language Models through data curation
• Identify issues, implement solutions, and report tooling bugs

Skills

Dutch English proficiency

Analytical skills

Attention to detail

Language data annotation

Adaptability

Education

MBO level 4 or equivalent

High school diploma

Job description

Amazon in The Hague is seeking an AI Data Trainer to create, annotate, and review training data across text, image and video. You will write texts in multiple styles following guidelines and perform quality checks to improve language models.

You will collaborate with scientists and engineers to refine guidelines, analyze data, identify issues, and propose solutions. The role focuses on data and tool improvements while handling sensitive content.