Hi, I'm Nate!

Needed accurate real-time detection of basketball makes and misses from user-uploaded video footage with varying angles and lighting conditions.

Could use traditional computer vision with OpenCV, a pre-trained model, or train a custom YOLO model on basketball-specific data.

Trained a custom YOLO model from Ultralytics on a basketball shooting dataset.

A custom-trained model provides significantly higher accuracy for this specific use case compared to generic models. YOLO's speed allows for efficient video processing while maintaining the precision needed to distinguish between makes and misses.

Need to block distracting websites during focused work sessions.

Could use browser extensions or a standalone desktop app.

Built a web-based solution with browser integration.

A web-based solution allows for cross-platform compatibility and easier updates without requiring users to install browser extensions.

OpenAI API calls can take 10-30 seconds to complete. A standard synchronous API route in Next.js would time out, providing a poor user experience.

I considered using webhooks or a background job queue like BullMQ with Redis.

I chose to implement a serverless-friendly queue system using Inngest.

While setting up a separate Redis instance for BullMQ would work, Inngest provides a serverless-friendly solution that integrates seamlessly with Next.js and handles retries, scheduling, and monitoring out of the box. This reduces infrastructure complexity while maintaining reliability.

Need to handle real-time updates across multiple clients.

Could use polling, Server-Sent Events, or WebSockets.

Implemented WebSocket connections using Socket.io.

WebSockets provide true bidirectional real-time communication with lower latency than polling and better browser support than SSE for this use case.

Need to handle real-time updates across multiple clients.

Could use polling, Server-Sent Events, or WebSockets.

Implemented WebSocket connections using Socket.io.

WebSockets provide true bidirectional real-time communication with lower latency than polling and better browser support than SSE for this use case.

Need to handle real-time updates across multiple clients.

Could use polling, Server-Sent Events, or WebSockets.

Implemented WebSocket connections using Socket.io.

WebSockets provide true bidirectional real-time communication with lower latency than polling and better browser support than SSE for this use case.