All work

SportsVisio, 2022–2025

Pro-level stats for every league

Advanced stats and highlights used to belong to major leagues. I owned the computer vision product that brings them to local leagues and players, from a one-hour game upload to stats, highlights and the apps.

The system

  1. Game upload
  2. Parallel processing
  3. YOLO inference
  4. Stats and highlights
  5. Mobile and web apps

Disciplines I coordinated

Design, engineering, data, ML & evals, compliance, business

  • 6,000+paying users
  • 39%retention
  • 2×Product of the Year finalist, 2023 and 2024

Context

SportsVisio uses computer vision to turn game video into the stats and highlights that used to be available only to major leagues. The users are local players, coaches and smaller leagues, through a mobile app and a coach web app.

My role

AI product manager for the computer vision product, end to end, from November 2022 to March 2025.

What I did

  • Owned the CV pipeline. Video ingestion, model inference with YOLO, stat generation, auto-generated highlights and delivery to the apps. One-hour game uploads are processed in parallel on AWS.
  • Evals with the ML team. Edge-case testing for lighting, occlusion and player overlap, feeding recurring model-improvement cycles.
  • Reliability you can see. Regression tests, latency and error tracking, and drift detection dashboards.
  • Privacy-safe onboarding. Time-bound links move users from web to mobile and attach the right permissions automatically, to meet league privacy requirements.
  • From one sport to two. Expanded from basketball to volleyball with sport-specific scoring logic, stats and UI.
  • Built-in sharing. AR effects and performance-based posters that players share on social.
  • Operations. Release plans, metrics and the customer center.

Outcome

6,000+ paying users with 39% retention. Product of the Year finalist at the New Hampshire Tech Alliance in 2023, for AI basketball analytics, and in 2024, for the volleyball expansion.

Stack and tools

  • YOLO
  • Python
  • FastAPI
  • Node.js
  • React
  • Next.js
  • TypeScript
  • Ruby on Rails
  • PostgreSQL
  • MongoDB
  • AWS
  • GCP
  • Firebase
  • Figma