
Algorithms that look perfectly efficient on paper often crash or cause severe performance bottlenecks when running on real physical hardware. Also, educational code is usually tangled with graphical interfaces (GUIs) that freeze the system during heavy workloads. OptiCode QA solves this gap between theory and practice through an automated testing pipeline powered by IBM Cloud. First, our tool "purifies" the code, stripping away any visual rendering to run the algorithm in a clean, pure-logic state. Second, we run a local Python stress test that hits the code with massive data loads (up to N=5000). We use precise telemetry to measure exact execution time in microseconds and peak RAM usage in bytes. Finally, we combine these real hardware metrics with the theoretical Big-O models generated by the IBM Bob IDE. The core of our project is the integration with IBM watsonx.ai. Using the Hackathon Sandbox, our system automatically sends the local telemetry data and the theoretical context to the IBM Granite 8B Code model. This AI acts as a smart code reviewer, cross-referencing the real physical execution against the theoretical limits to output a final compliance report. OptiCode QA ensures that code is rigorously tested for stability, memory limits, and performance bottlenecks before it ever reaches a production environment.
17 May 2026