1 of 1 Permanent Benchmarking Jobs in Maidenhead

Help define the evaluation strategy and methodology for AI capabilities across the company – automated metrics, human‐in‐the‐loop workflows, test set management, and benchmarking – and establish the quality standards other teams build against. Build evaluation frameworks and developer tooling robust enough for production yet simple enough for non‐specialist … such as COMET, chrF++, BLEU, MetricX, and MQM‐style human evaluation. Experience with experimentation and observability tooling, data/test‐set versioning, and rigorous benchmarking workflows. Established practice in AI governance and documentation - model cards, system cards, reproducibility, and responsible‐AI considerations - at an organizational level. Broad familiarity with ...