Giskard

Giskard is a company that specializes in providing an open-source solution for AI quality assurance, aiming to address common risks such as biases, performance issues, and security vulnerabilities in AI models. With its tools, users can improve their AI models with fewer than 10 lines of code, supporting a wide range of models from tabular to Large Language Models (LLMs). Giskard’s offerings include the ML Testing Library, an open-source Python library designed to detect hidden vulnerabilities in ML models and LLMs automatically. Additionally, the company offers the AI Quality Hub, an enterprise-ready platform featuring visual testing, explainability, and debugging dashboards for collaborative AI quality assurance at scale. Giskard also introduced LLMon (beta), a tool for production monitoring of LLM-based applications, providing real-time alerts for issues like hallucinations, incorrect responses, and toxicity. The company is trusted by notable clients such as L’Oréal, AXA, iAdvize, and Banque de France and has been listed by Gartner in the AI Trust, Risk, and Security categories. Giskard aims to unify AI testing practices using standard methodologies to ensure optimal model deployment and compliance with AI regulations. It also fosters a community platform for sharing best practices, creating new tests, and shaping the future of AI safety standards.