Composo
.avif)
Composo provides an API for deterministic evals / quality scoring of AI agents & outputs.
About the Tech
They've cracked the code on deterministic AI evaluation. While everyone else is using unreliable LLM-as-judge with 30%+ variance, the Composo team built purpose-built generative reward models that deliver 92% accuracy with <1% variance. Teams can finally trust their evaluation metrics. Plus, you just write one sentence to create any custom criteria - no complex prompt engineering needed.
It's a complete paradigm shift from hoping your AI works to knowing it works.
About the Team
Deep Enterprise AI Experience: The founders combine strategy/product expertise (Dr. Sebastian Fox - ex-McKinsey/QuantumBlack) with technical engineering depth (Luke Markham - ex-Graphcore ML unicorn), both with 5+ years building large-scale AI systems for Fortune 500 companies.
Proven Track Record: They've worked with leaders at companies like Palantir, Tesla, Accenture, UBS, and other major enterprises, giving them real-world experience shipping production AI applications at scale.
Elite Background: Both founders have Oxford University degrees and experience at top-tier firms, bringing a unique blend of academic rigor and practical enterprise deployment expertise.
Composo would not exist without EWOR and their support both at the very beginning of the company, and continuing to this day post-raise and 2 years in.
More from EWOR
