Company Name: Patronus AI
Business Focus: Automated evaluation and security platform for the safe usage of large language models (LLMs)
Funding Details: Patronus AI secured $3M in seed funding.
Purpose of Funding:
The funding will be directed towards enhancing the capabilities of their platform and furthering their mission to provide secure LLM evaluations.
Patronus AI is the brainchild of machine learning authorities Anand Kannappan and Rebecca Qian. The company's primary objective is to equip enterprise development teams with proprietary AI tools that can efficiently assess model performance, formulate adversarial test cases, benchmark models, and much more. By automating and scaling the manual and typically expensive model evaluation techniques that enterprises commonly employ today, Patronus AI paves the way for organizations to implement LLMs while substantially reducing the potential risks of model failures and incorrect outputs.
The platform provided by Patronus AI streamlines several processes:
- Scoring: It evaluates model performance in real-world scenarios focusing on essential metrics such as hallucinations and safety.
- Test generation: The system can automatically generate adversarial test suites on a large scale.
- Benchmarking: It facilitates model comparisons, assisting customers in identifying the optimal model tailored to specific use cases.
Several leading AI enterprises, like Cohere, Nomic AI, and Naologic, are early adopters of the Patronus AI platform. Moreover, several renowned companies from conventional sectors, particularly financial services, have plans to test Patronus AI in the forthcoming months.