Data Scientist

We're looking for a Data Scientist to embed with our service delivery and client teams and turn large, messy operational data into decisions. You'll work at the intersection of data quality, human annotation, and model evaluation, measuring rater agreement, surfacing quality signals, building predictive and diagnostic models, and translating findings into recommendations that leadership and clients act on. This is a hands-on, business-facing role for someone as comfortable writing SQL and Python as they are presenting a trade-off to a non-technical stakeholder.

Responsibilities:

• Analyse rater and annotation data to measure quality, consistency, and agreement, including inter-rater reliability, precision/recall trade-offs, and temporal validation of task performance.
• Build and validate statistical and machine learning models to predict quality issues, detect anomalies, and support workforce and capacity planning.
• Design and evaluate metrics and experiments (A/B tests, sampling strategies, audit frameworks) to assess data and model quality for client programs.
• Partner with delivery leads, program managers, and client stakeholders to frame business questions, scope analyses, and deliver clear, actionable insights.
• Work with large datasets in BigQuery and GCP, writing efficient, well-tested SQL and Python for analysis and pipelines.
• Communicate results to technical and non-technical audiences through clear visualisations, concise narratives, and executive-ready summaries.
• Collaborate with data engineering and BI teams to productionize recurring analyses and move ad hoc work toward reusable, industrialised reporting.

Requirements:

• 4+ years of experience in data science and quantitative analysis.
• Strong applied statistics: hypothesis testing, sampling, regression, and evaluation metrics (precision, recall, F1 agreement/reliability measures).
• Proficiency in Python (pandas, scikit-learn, NumPy) and strong SQL.
• Experience working with large datasets and cloud data warehouses (BigQuery or equivalent).
• Ability to translate ambiguous business problems into structured analyses and communicate findings clearly to stakeholders.
• Bachelor's degree in a quantitative field (Statistics, Computer Science, Mathematics, Economics, Engineering) or equivalent practical experience.

Nice to Have:

• Experience with human-in-the-loop data, annotation quality, or crowd/rater workflows.
• Familiarity with GCP tooling (Dataflow, Cloud Run, Vertex AI) and modern data stack tools (dbt, Looker).
• Exposure to LLM/GenAI evaluation, model benchmarking, or data-for-AI programs.
• Experience with experimentation and causal inference.
• Master's degree in a quantitative discipline.

Skills

python, sql, statistical, statistics