California-based automated AI model testing startup, RagaAI, emerged from stealth today with $4.7 million in seed funding led by pi Ventures. Participating investors in the round included Anorak Ventures, TenOneTen Ventures, Arka Ventures, Mana Ventures, and Exfinity Venture Partners.
The capital will advance research and development, improve their AI testing tools and form strategic partnerships.
Founded in 2022 by tech veteran Gaurav Agarwal, RagaAI’s breakthrough foundation models uses automation to detect AI model issues, diagnose and fix them instantly. They reduce 90% of the risks while accelerating AI development by more than 3x.
The platform already offers over 300 different tests to help users triage the issue down to its root cause. It’s able to identify issues as varied as data drift, edge case detection, poor data labelling quality, bias in the data, lack of model robustness or adversarial attacks. They’re multimodal, supporting LLMs, images/videos, 3D, audio, NLP and structured data.
The AI market is expected to reach $2 trillion by 2030. Generative AI is forecast to add $10.3 trillion to the global economy by 2038. RagaAI believes 25% of that market size spend will be targeted towards ensuring AI is safe and reliable.
“At Ola and NVIDIA, I saw the significant consequences of AI failures due to lack of comprehensive testing,” explained Agarwal. “Our foundation model “RagaAI DNA” is already solving this problem across large Fortune 500 companies. Our vision is to liberate AI from the constraints of human interventions.”
RagaAI has proven its utility across various sectors. For an e-commerce client, it refined a chatbot, reducing response errors by identifying and fixing hallucinations. In the automotive industry, it enhanced vehicle detection in low-light conditions by simulating various lighting scenarios, improving accuracy and potentially preventing accidents.
RagaAI prioritizes enterprise needs and data privacy. They hold SOC2 Type2, ISO 27001, HIPAA, GDPR, and CCPA certifications. Their technology is adaptable for private clouds or on-premise deployment. The platform is available through a UI and Python interface.
“Guided by our core values, we are committed to pushing the boundaries of automated AI issue detection and automated root cause analysis,” added Agarwal.
