Objective

Build a system that analyzes indoor rowing form in real time and provides feedback to the user while they row.

Rowing Form

In rowing, one complete back-and-forth motion is called a stroke. Each stroke consists of four phases:

• Catch — front position (knees bent, arms extended, torso slightly forward, shins close to vertical)
• Drive — movement from front to back (legs push first, then torso swings back, then arms pull)
• Finish — end position at the back (legs extended, torso slightly leaning back, handle near chest)
• Recovery — movement from back to front (arms extend, torso moves forward, then knees bend)

These phases repeat continuously as part of each stroke.

Also, maintain a neutral back throughout the stroke (avoid rounding or excessive arching).

Expectations

The entire solution runs on-device on an iPhone and works offline.

The solution may use cloud services, but raw video or image data must not leave the device (please clearly describe what runs locally vs remotely).

Solutions closer to the preferred setup will be given higher weight during evaluation.

Deliverables

• An iOS app testable via a TestFlight public link that runs on iPhone 16 Pro
• Demo video
• Brief documented explanation of your approach
• Source code

Evaluation Criteria

• Effectiveness of form detection (accuracy and consistency in identifying issues)
• Quality of feedback (how correctly and clearly issues are communicated while rowing)
• Real-time performance (responsiveness during actual rowing)
• Craftsmanship (attention to detail and user experience)

Exceptional submissions may be considered for collaboration even if not selected as the winner.

Interested? Reach out at [email protected]