Rockfish Data
Synthetic data for testing and evaluating AI models and agents before they reach production.
Private company- Sector
- AI
- Based in
- San Ramon, United States
- Founded
- 2022
- Invested
- Jul 2023
Rockfish Data builds a platform that generates synthetic data from schemas or sample datasets while keeping the statistical and time-based structure of the original. Its current focus is AI evaluation: producing labelled, domain-specific test sets, including rare events and anomalies that production data seldom contains, so teams can find where models and AI agents fail before they ship.
The platform runs inside a customer's own environment, whether a private cloud, on-premises or as a Snowflake Native App, so real data does not leave the organisation. It has been used on operational data such as financial transactions, cybersecurity telemetry and supply chains, with customers including streaming analytics company Conviva and US government agencies.
Rockfish was founded in 2022 by Muckai Girish together with Carnegie Mellon University researchers Vyas Sekar and Giulia Fanti, and builds on their research at CMU. It is based in California's Bay Area, with a presence in Pittsburgh.
