
Helping AI engineers to evaluate generative models up to 180x faster.
Core product: Adaptive LLM evaluation platform using Item Response Theory / computerized adaptive testing and a Scorebook framework
Headquarters: Cambridge, United Kingdom
Founding origin: University of Cambridge research (Prof. Nigel Collier)
Team size (reported): 4 employees
Recent funding: £2.2M pre-seed (announced Sep 2025)
Evaluation and safety testing of large language models
2024
AI / Machine Learning
£2.2M
Participating investors reported: Cambridge Enterprise Ventures, Parkwalk Advisors, Fund F, Vento Ventures and angel investors (Ventures Together)
“Led by an institutional early-stage investor (Twin Path) with participation from university-linked and specialist investors (Cambridge Enterprise Ventures, Parkwalk Advisors)”