SchedSim Lens was inspired by a simple goal: make operating systems concepts easy to understand. Deadlocks and race conditions are among the hardest OS topics, because they depend on timing, hide when you read code, and rarely show up when you run it. We wanted students to learn them by seeing them, and to fix them in the programming language they are most comfortable with, from Python to C++ to Java.

Upload a photo, screenshot, or snippet of multithreaded code. Gemma 4 uses vision and structured JSON output to extract each thread’s lock and variable operations into one language-neutral model. An editable review table with a confidence score lets you correct any misread line. A deterministic engine then searches thread interleavings to find deadlocks and race conditions, and animates them as a calm timing diagram with a wait-for graph synced to the code.

Gemma explains the bug in simple language and proposes a fix written in your own language. The fix goes back through the same pipeline, and the app shows “Fix verified” only if the engine confirms it. The model extracts; the engine proves.

All model calls go through one adapter, so the same prompts run on the Gemini API or a self-hosted open-weight Gemma, keeping student code private. The project is open source under the Apache 2.0 license.