Hiring a data scientist for a machine-learning engineering role, or the reverse, is one of the more expensive avoidable mistakes in early-career technical hiring.
Live from our candidate database
Counted at render time and refreshed every six hours. This is a supply-side view of one platform, not a national labour-market statistic.
Skills present in this pool
The dividing line is production
Data science ends at inference and insight. Machine-learning engineering begins at serving one — latency, monitoring, versioning, the behaviour of a model that degrades quietly at three in the morning.
Screen ML engineers on engineering
Solid Python beyond notebooks, comfort with versioning and reproducibility, and the ability to reason about a model that has started drifting. Modelling depth is secondary for this role and is often over-weighted.
Screen data scientists on reasoning
Why this model, how did you know it worked, what would break it. Deployment experience is a bonus rather than the requirement.
Say which one you want in the brief
Most early-career candidates will present themselves as either, because both titles are aspirational. Being explicit in the requirement saves everyone a round of interviews.
Key takeaways
- Production is the dividing line: inference versus serving.
- Screen ML engineers on Python engineering, reproducibility and drift reasoning.
- Screen data scientists on model justification and evaluation discipline.
- State which role you want in the brief — candidates will otherwise present as both.
Questions
Can one person do both at entry level?+
Rarely well. Expecting both from a graduate hire usually produces someone weak at each, and the expectation is on you rather than on the candidate.
Which is harder to hire?+
ML engineering, consistently — the production-facing skills are almost never taught at undergraduate level.