There’s no universal definition of a “data scientist,” which is why hiring one can be difficult. The best start with clarity — define the *problem*, not the *person*.

Are you trying to build predictive models, automate reporting, or understand customer behaviour? Each goal calls for a different mix of statistics, engineering, and business context. Get that wrong, and even great candidates will struggle.

The right hire depends on maturity. Early-stage teams often need generalists who can code, visualise, and communicate. Mature teams need specialists who can scale pipelines and productionise models.

When I work with clients, I help translate these differences into job specs that attract the right talent the first time.