Pretraining Data Researcher: Dataset Design & Quality

Microsoft; City of Westminster; On-site; GBP 75,000 - 123,000; Full time

Job summary

Microsoft AI is seeking Researchers and Research Engineers to join the Pre‑training Data team in the United Kingdom. You will design and curate high‑quality datasets, develop scalable data pipelines, and assess data quality to advance frontier AI models.

Collaboration with Safety, Ethics, and Governance ensures responsible AI practices. This role emphasizes data-driven model improvement over academic publishing.

Qualifications

• Bachelor’s or Master’s in a technical field with strong data focus.
• Proficiency in Python and data libraries (Pandas, NumPy).
• Experience with large-scale, unstructured data and data engineering.
• Strong knowledge of statistics and exploratory data analysis (EDA).
• Experience with Spark, Ray, or Apache Beam.
• Ability to communicate technical findings to research and product teams.

Responsibilities

• Create high‑quality datasets for training and evaluation; run experiments on new datasets to assess impact.
• Develop and maintain scalable data pipelines for ingestion, processing, filtering, and annotation.
• Analyze real‑world datasets to assess quality, diversity, relevance, and identify improvements.
• Build lightweight tools and workflows for dataset auditing, visualization, and versioning.
• Collaborate with Safety, Ethics, and Governance teams to ensure datasets meet standards for quality, privacy, and responsible AI practices.

Skills

Python

Data libraries

Statistics

Exploratory data analysis

Data processing frameworks

Communication of findings

Large-scale data handling

Education

Bachelor's degree in AI/CS/Data Science/Statistics/Physics/Engineering

Master's degree in AI/CS/Data Science/Statistics/Physics/Engineering

Equivalent experience

Tools

Spark

Ray

Apache Beam

Job description

Microsoft AI is seeking Researchers and Research Engineers to join the Pre‑training Data team in the United Kingdom. You will design and curate high‑quality datasets, develop scalable data pipelines, and assess data quality to advance frontier AI models.

Collaboration with Safety, Ethics, and Governance ensures responsible AI practices. This role emphasizes data-driven model improvement over academic publishing.