CNN Server MVP

A 3D convolutional neural network with:

• A POSIX asynchronous HTTP server with a pthread thread pool
• A ZeroMQ-based in-memory queue for dispatching image classification requests
• A worker process that consumes requests, runs CNN inference, and logs results
• A setup for horizontal scaling and easy integration with a load balancer (e.g., NGINX)

---

Features

• HTTP GET /predict?img=/path/to/image.jpg endpoint
• Basic HTTP parsing and response
• Thread pool for concurrent request handling
• ZeroMQ PUSH/PULL queue for decoupled request dispatch
• Simple file path validation