An Xray with a bounding box highlighting an abdominal mass

The first abdominal mass located by the mass detection model. 🎉

RapidRead: State of the Art AI Radiology System for Pets @ MARS

This AI system looks at X-rays for dogs and cats and can detect the presence of 53+ clinical findings such as cardiac issues, masses, and intestinal obstructions.

Since joining the team in 2021 my contributions to the project have been:

  • Developed and deployed a Mass Detection AI model that detects masses in veterinary x-rays.

  • Cross-functional leadership; working with data scientists, product managers, stakeholders, and senior leadership.

  • Experimenting with state-of-the-art approaches for model training and evaluation.

  • Deploying new models into production.

  • Developing the “RapidRead Management Tool“ that allows our stakeholders to manage AI systems, turning stakeholders into developers.

  • Implementing Agile processes, workflows, and ceremonies.

Technology Used:

  • Python 🐍

  • Pytorch🔥

  • Convolutional Neural Networks 🤖

  • FiftyOne

  • MLflow📈

  • FastAPI 👨‍🍳

  • Docker Containers 📦

  • Azure ☁

My Line Manager, Michael Fitzke, presenting our project at the Pytorch Conference in 2022.

MARS Chief Digital Officer, Sandeep Dadlani, calling out RapidRead as one of his favorite AI projects in an interview with Forbes.

Next

AI Classifier for Healthy Coral Reefs: An AI Hackathon Challenge