trismik
Research led
38 +
years of combined research experience
200 +
published research papers including ACL, EMNLP, NEURIPS, and AAAI
12,000 +
author citations
We believe LLMs should be useful and safe for everyone. As a Cambridge spin-out, we've contributed core NLP methods to the research community. At Trismik, we focus on efficient real-world evaluations - combining human and model judgements to deliver scalable, automated testing
Founding team
Rebekka Mikkola (CEO)
Ex Salesforce
Nigel Collier (CSO)
University of Cambridge
FAQs
What's our back story?
We are an early-stage spinout founded by tech geeks from Cambridge, Salesforce and Amazon. Backed by 38+ years of research, we’re on a journey to bring the highest scientific standards to AI evaluation.
What sets us apart?
We use adversarial testing to uncover model weaknesses with fewer queries - helping teams find failure modes standard benchmarks miss.
How do we deliver up to 180x faster evaluations?
We reduce evals from 17K+ queries to under 80 by focusing on high-signal test items. This means significantly faster, cheaper runs - without losing diagnostic power.
Featured in
