Research Engineer, ML Evaluation & Data

Harvey; Greater London; On-site; GBP 90,000 - 150,000; Full time

Job summary

Harvey is building data and evaluation systems to power post-training and benchmarking for its agentic AI platform. You’ll own synthetic data pipelines, evaluation design, tasks and grading behind training environments, and the systems turning expert feedback into signals and rewards.

This is a high-bar, impact-first research engineering role. You’ll join a small team of engineers and researchers, turning model failures into measurable improvements while collaborating with researchers and legal

Qualifications

• Experience building data pipelines and ML evaluation systems.
• Strong Python skills with ownership of production-quality code.
• Rigorous measurement approach: experimental design and metric interpretation.

Responsibilities

• Build synthetic data pipelines for training and evaluation samples.
• Design evaluations and verifiers reflecting real legal work and calibrate against expert judgment.
• Build and maintain capability benchmarks like the Legal Agent Benchmark (LAB).
• Collaborate with legal domain experts to define quality metrics and rubrics.
• Own training and evaluation data quality, provenance, and safeguards against contamination.
• Close the loop from product usage to new evaluations and training data enhancements.
• Partner with researchers to analyze model failures and guide training decisions.

Skills

Data pipelines

ML evaluation systems

Python

Experiment design

Job description

Harvey is building data and evaluation systems to power post-training and benchmarking for its agentic AI platform. You’ll own synthetic data pipelines, evaluation design, tasks and grading behind training environments, and the systems turning expert feedback into signals and rewards.

This is a high-bar, impact-first research engineering role. You’ll join a small team of engineers and researchers, turning model failures into measurable improvements while collaborating with researchers and legal