RadMate AI

Overview

RadMate AI was a San Francisco-based healthcare AI startup founded in 2023 by Mohamed Khalifa and Adam Skrocki, two Cornell Computer Science alumni accepted into Y Combinator's Winter 2024 batch.^[1] The company built an AI copilot for radiologists — a workflow tool that listened to dictated findings and used large language models to generate structured, proofread radiology reports, reducing the volume of words radiologists needed to speak and the number of clicks required to produce a final report.^[2]

RadMate AI failed because it entered a capital-intensive, compliance-heavy enterprise market with $500K in pre-seed funding and a two-person team — a structural mismatch that made it nearly impossible to survive the 12–24 month sales cycles required to reach paying hospital customers.^[3] A direct competitor, Rad AI, had already locked up distribution across a third of all U.S. health systems before RadMate AI presented at Demo Day.^[4]

No formal shutdown announcement has been made. By 2025, the company's website returned no content, its LinkedIn page had accumulated only 250 followers, and CEO Mohamed Khalifa's LinkedIn title had shifted to "Operator & Advisor | Y Combinator" — indirect signals that active operations had ceased.^[5][6] No seed round was ever announced following YC Demo Day, and no acquirer or acqui-hire has been publicly documented.

Image 1 / 2

Founding Story

Mohamed Khalifa and Adam Skrocki met as Computer Science students at Cornell University.^[7] Their paths after graduation gave them complementary and directly relevant experience for what they would eventually build together.

Khalifa joined Amazon Web Services as a Customer Solutions Manager before moving to PathAI, a computational pathology company, where he worked as a Software Engineer building an image viewer platform designed to improve the workflow of pathologists.^[8] That role gave him a specific lens: he had seen firsthand how AI could be layered onto a clinical imaging workflow to reduce friction for specialists — not by replacing the clinician's judgment, but by automating the documentation burden around it.

Skrocki took a different route, joining Palantir as a software engineer on the AI Platform team, including its initial launch.^[9] His experience building enterprise AI infrastructure gave him a technical foundation for deploying LLM-based tools in complex organizational environments. But the founding motivation was more personal: Skrocki's father is a radiologist, and the daily friction his father experienced — dictating lengthy reports, correcting errors, navigating clunky software — became the product brief.^[10]

The insight the founders brought to Y Combinator was specific: radiology reporting was a documentation problem masquerading as a clinical problem. Radiologists were already expert image readers. What slowed them down was the translation of visual findings into structured written reports — a process that required dictating more words than necessary, manually formatting output, and then catching errors before submission. Khalifa's PathAI experience had shown him that this exact workflow bottleneck existed in pathology; radiology, with its higher report volumes and more standardized formatting conventions, was a larger and arguably more tractable version of the same problem.

The company was accepted into Y Combinator's Winter 2024 batch and publicly launched on February 5, 2024.^[11] No major product pivot has been documented — the company appears to have pursued its original thesis from founding through Demo Day without a significant change in direction.