Data Science Intern

About the role

Make data dependable, understandable and useful for decisions. Build analyses that connect a clear question to sound measurements and explainable findings.

What you’ll work on

• Define useful questions and metrics before selecting an analysis method.
• Clean and reconcile datasets, checking missing values, duplicates, outliers and inconsistent definitions.
• Use SQL and Python to explore patterns and create reproducible reports.
• Analyse workflow quality or experiment results with attention to uncertainty and sampling bias.
• Present findings, assumptions and limitations clearly enough to support a practical decision.

Core skills

• SQL joins, aggregations and window functions, with an understanding of relational data.
• Python and pandas for cleaning, transformation and exploratory analysis.
• Descriptive statistics, probability, sampling and the distinction between correlation and causation.
• Clear visualisation using tools such as Matplotlib, Seaborn or a BI platform.
• Data-quality checks, reproducible notebooks and concise written interpretation of results.

Useful additional experience

Helpful, but not required.

• Confidence intervals, hypothesis testing or experiment design.
• Document-quality benchmarks, text analysis or labelled evaluation datasets.
• Versioned data pipelines or dashboards with documented metric definitions.

Where this work connects

Arvi’s document intelligence and PDF-to-HTML workflows create questions about output quality and review effort. This track would develop evaluation datasets and analyses for those questions, alongside scoped product and service work.

Show us your work

Share an analysis with its question, data source, cleaning steps, charts and recommendation. Explain one limitation that changes how the result should be interpreted; a public dataset or coursework project is welcome.

Tell us what you contributed, what you learned and what you would improve. Personal projects and coursework are welcome.