Empirical Security

AI models that help enterprises detect and mitigate risks in LLM-generated content

Seed · $9M · 2025-06

Open-source software that helps enterprises build, monitor and improve LLM applications in production

Seed · $7M · 2025-08

Platform for companies to customize existing AI models with their own data and brand voice

Seed · $7M · 2025-02

ML architecture combined with formal logic to generate rigorously verified control software for safety-critical fields such as automotive, aerospace, defense and robotics, aiming to eliminate the "hallucinations" and errors of conventional LLMs in control code

Pre-Seed · $3M · 2025-12

Online RL-based optimization of AI model training (Paramorph) and large-scale training observability (Metrana)

Undisclosed · $2M · 2025-04

Automatically evaluates frontier LLMs' performance and cost in a company's specific use cases and recommends model selection

Undisclosed · $2M · 2023-10

Enterprise AI reliability: builds and runs enterprise AI agent systems on in-house platform Nora, using an in-house ~27B-parameter hallucination detection model to check each agent response, trace root causes and close the loop on fixes; small models to cut inference costs