Computer Vision Platform for Collegiate Football Recruiting

A national sports company asked us to help them develop an analytics platform to calculate the speed of football players from game film. The speed metrics obtained would provide valuable supplemental data to the company's existing collegiate recruiting database customers.

Our team designed and built a two-part solution including:

(1) a computer vision backend to process videos and calculate the speed of players, and

(2) an intuitive web application to allow analysts to upload videos and review resulting metrics.
‍

Our stack for this platform included Python for the vast majority of the computer vision and deep learning loads, and Javascript + React for the frontend web application. We deployed the application in Google Cloud Platform.

(1) Strategic research - what's currently out there? any competing products?

(2) Design for prototype - what is the fastest way to get to a working model that proves technical feasibility?

(3) Rapid prototyping - we spun up a basic proof of concept in under 2 weeks

(4) Design for MVP - what does a minimum viable product (MVP) look like?

(5) MVP Engineering - we built and deployed a sales-ready MVP in 3 months

(6) Beta - we worked with the client to launch a beta and show off the platform

(7) Release Candidate & V1 - we planned a roll out of a complete production release

(8) V1 Engineering - iterating on the MVP, we built out more complete and advanced functionality

(9) Launch & Training - we launched a proper v1.0 release and conducted hands-on training with client users

(10+) We are still engaged with the client and working toward v2.0