TrainWise AI

Video Demo: https://www.youtube.com/watch?v=A2UBi_SAaDg

TrainWise AI is a workout tracker app which uses Google’s PoseNet AI model to count and track your reps for you.

The app was build with React Native, and uses Express.js backend APIs with Node.js.

The app uses a locally saved PoseNet AI/ML model, which takes in the video feed of the user and turns it into a set of points that represent various joints of the body.

Using these points, and how they change over time, we can count the number of reps being performed by a user.

As technical lead, my responsibility was to:

• Ensure successful integration between frontend and backend by advising respective team leads and assisting in development,
• Achieve alignment between system architecture and business objectives by researching and designing the end-to-end system, and
• Enable team members to focus on development by initializing and maintaining necessary development environments, processes, and documentation, as well as reviewing implemented code and enforcing necessary coding standards.

Opening the app takes us into a home page.

We can see past workouts on the workouts page.

The current workout page has workout templates to choose from, or you can start a new workout.

Here, we can add and edit sets for any given workout.

Clicking the camera tracker button on an exercise will open the AI camera tracker.

After setting up your phone on the floor, you can start performing the exercise, and the reps will automatically be counted for you.