Physician-led, AI-powered, continuously refined

EndlessMed was created and is overseen by a practicing consultant rheumatologist who is board-certified by the American Board of Internal Medicine in both Internal Medicine and Rheumatology and has remained current through every recertification cycle since initial certification.

He passed Step 1, Step 2 CK, and his ABIM board examinations on the first attempt. EndlessMed began, in part, as a project to help his own children prepare for Step 2 CK — and grew into a platform designed to help other medical students prepare with the same emphasis on clinical reasoning, evidence, and accuracy.

AI generates the questions. It doesn't get the final word.

Every EndlessMed question is generated using AI, but generation is only the beginning.

Questions go through a multi-step quality process that includes structural and consistency checks followed by an independent AI review focused on medical accuracy and alignment with current clinical guidance. Questions that do not pass the required checks are not presented to students as verified content.

Students are also part of the quality process. Every question can be flagged using “Report this question.” A reported question can be taken out of rotation for further review, and a physician makes the final decision when a potential problem requires human review. Questions can then be returned to the bank, kept out of rotation for further assessment, or removed.

No AI system is treated as infallible — and neither is the question bank.

Built with multiple AI systems, not just one

AI is also deeply involved in the development and continuous improvement of EndlessMed itself.

The platform has been developed using a multi-model workflow involving six different AI systems and agents, each contributing to different aspects of development, review, testing, analysis, and improvement. This allows ideas and implementations to be challenged from multiple directions rather than relying on the output of a single AI system.

Since development began, EndlessMed has gone through nearly 800 individual iterations and modifications, with physician oversight throughout.

And that process continues. The goal isn't to claim that EndlessMed is finished or perfect. It is to keep identifying weaknesses, testing assumptions, and making the platform better.

Physician oversight doesn't mean pretending to know every specialty

Step 2 CK covers a far broader range of medicine than Internal Medicine and Rheumatology.

Our founder's board certifications are specifically in those two fields. His role is therefore not presented as personal specialist verification of every obstetrics, pediatrics, surgery, psychiatry, or other specialty question.

Instead, he sets the clinical standard to which the platform and its verification process are held, reviews issues that require human judgment, and oversees how EndlessMed approaches medical accuracy and clinical reasoning.

More than a question bank

EndlessMed was designed not only to tell you what you're getting wrong, but to help identify why.

As you answer questions, the platform looks for patterns in your performance — for example, whether difficulty appears more consistent with a knowledge gap, integrating multiple clinical findings, or particular types of exam tasks. It also tracks how those patterns change as you continue practicing.

The goal is more than another collection of practice questions. EndlessMed is a continuously evolving learning system built around clinical reasoning, evidence-based medicine, transparent use of AI, and physician oversight.