Yrikka

Yrikka AI Inc. is a New York-based company specializing in AI risk management, particularly for visual AI systems in high-stakes sectors. Founded in 2021 by CEO Dr. Kia Khezeli and CTO John Kalantari, the company provides tools for organizations to ensure their AI models are safe, reliable, and aligned with human values. The founders' backgrounds are deeply rooted in machine learning and AI safety, with Khezeli holding a Ph.D. in Electrical and Computer Engineering from Cornell with a focus on AI safety, and Kalantari holding a Ph.D. in Biomedical Engineering. Their collective experience at institutions like the Mayo Clinic, NASA, Google, and Intel inspired them to address the lack of tools for building trust in AI systems.

The company's core offering is a platform for "contextual red-teaming" of visual AI models. This process involves systematically testing AI systems to uncover vulnerabilities and weaknesses by simulating real-world attacks and edge cases. Yrikka automates this traditionally manual and time-consuming process, using AI agents to assess how a model performs in the specific context of its deployment, rather than in abstract settings. This approach significantly reduces the time needed to validate new models from months to minutes. The platform is delivered via an Application Programming Interface (API), starting with its "APEX" (Adversarial Probing and Evaluation of Exploits) API. The business model centers on providing this API access, enabling clients to automate risk assessment, monitor for issues like data drift and adversarial attacks, and continuously fine-tune their AI assets.

Yrikka targets industries where AI failure has serious consequences, including defense, healthcare, autonomous systems, industrial automation, and agriculture. A notable achievement is a $1.9 million contract with the U.S. Department of Defense to automate the testing and evaluation of its computer vision models. The company launched publicly after securing a $1.5 million pre-seed funding round co-led by Focal and Garuda Ventures.

Keywords: AI risk management, visual AI, computer vision, AI safety, red teaming, AI alignment, model validation, contextual red teaming, adversarial testing, AI evaluation, machine learning security, autonomous systems, defense technology, healthcare AI, industrial automation, AI reliability, data drift monitoring, human-aligned AI, APEX API, Kia Khezeli, John Kalantari