Applied MLE Engineer, North: Eval Systems & Product Quality

Cohere; Greater London; Hybrid; GBP 120,000 - 170,000; Full time

Benefits offered by this job

Lunch stipend

Health benefits

RRSP matching

Parental leave

Enrichment stipend

Education stipend

Vacation time

Job summary

Cohere is seeking a hands‑on Senior MLE minded engineer to own North's evaluation framework. You will define metrics across agent workflows, build robust evals from real usage, and ensure that measurement tracks product progress.

You will work closely with product and modelling teams to turn insights into model updates that improve North’s enterprise capabilities. You will translate customer needs into precise evaluation signals, maintain rigorous rubrics, and drive practical improvements for

Qualifications

• Strong applied MLE judgment and can reason about model behavior and eval validity.
• Experience building evals and feedback loops to improve product outcomes.
• Ability to translate messy signals from users into measurable evals for modeling teams.

Responsibilities

• Own the eval strategy for North: define what to measure across agent workflows and enterprise knowledge work.
• Build high-quality evals from user feedback, production failures, privacy-preserving logs, and customer needs.
• Create systems that turn learnings from users into evals, ensuring measurement keeps pace with the product.
• Extract clear insights from eval results and turn them into actionable recommendations for central modelling teams.

Skills

Eval design

Data curation

Model evaluation

Applied MLE judgment

Product awareness

Job description

Cohere is seeking a hands‑on Senior MLE minded engineer to own North's evaluation framework. You will define metrics across agent workflows, build robust evals from real usage, and ensure that measurement tracks product progress.

You will work closely with product and modelling teams to turn insights into model updates that improve North’s enterprise capabilities. You will translate customer needs into precise evaluation signals, maintain rigorous rubrics, and drive practical improvements for