MD.ai

MD.ai is a comprehensive platform designed to support medical deep learning through its FDA 510(K)-cleared MD.ai Viewer for diagnostic use. The platform enables users to create high-quality labeled training and validation datasets using web-based annotation tools optimized for medical applications. MD.ai supports API, Jupyter integration, and client libraries to facilitate model development, and offers federated learning for model training. Models can be run in the browser, on-premises, or in the cloud. Additionally, MD.ai offers a Reporting/Chat application powered by GPT-4 and other large language models (LLMs). The platform collaborates with various organizations to build large annotated datasets for medical AI applications, including data from The Cancer Genome Atlas Lung Adenocarcinoma collection. MD.ai provides cloud storage of anonymized medical images and supports real-time collaboration for medical deep learning projects. The platform allows easy export of annotations, images, and labels for training, and offers URL links for assigning cases to team members. MD.ai supports rapid prototyping of medical deep learning projects and leverages Google Cloud Platform and Cloud Healthcare API for creating annotated datasets and building algorithms. The software is configurable to meet HIPAA and other privacy and security requirements, providing an audit trail of all electronic events related to data within their system. MD.ai’s servers operate within ISO 27001 certified data centers and, where required, HIPAA certified data centers. The company maintains detailed audit logs of all internal systems and regularly reviews them for vulnerabilities, and conducts security training with all employees upon hire and regularly throughout employment.