AI research scientist
Create and test new methods with reproducible evidence.
The outcome this role owns
A defensible research contribution, with clear comparisons, reproducible experiments and an honest account of what remains unknown.
Four levels of maturity
Proposed framework · Evidence-based discussion · Not a validated assessment scale
Foundation
Reproduces a published result and identifies assumptions and missing implementation details.
Evidence to look for
A reproduction report including discrepancies
Pilot
Tests a focused hypothesis against credible baselines and records negative results.
Evidence to look for
An experiment log with appropriate baselines
Production
Leads a research project with reproducible experiments, ablations and independent scrutiny.
Evidence to look for
A reviewed contribution with code or a reproducibility package
Organisational
Shapes a research agenda, mentors researchers and decides which promising results warrant further resources.
Evidence to look for
An agenda and resource decision grounded in research results
A practical work sample
Provide a short paper claim and its reported benchmark result. Ask for alternative explanations, missing ablations and a feasible reproduction plan.
Use synthetic or public data, explain the evaluation criteria in advance and allow reasonable adjustments. Discuss the reasoning as well as the finished artifact.
What to question
Novelty is asserted without strong baselines, or only the best runs are reported.
Research talent needs different evidence
This release has no comparable research-talent measure. Enterprise adoption and ICT employment cannot identify frontier research capability, so no research-market ranking is provided.
For a research hire, review the relevant lab, subfield, publication quality and reproducibility of the work. A national business-adoption figure cannot answer those questions.
Framework sources
This guide is an original synthesis drawing on the sources below. The sources do not publish or endorse these role levels.