What it is
An open-source practice platform for system design and low-level design interviews. It walks users through interview-style stages, gives AI-generated feedback against a published rubric, and lets them bring their own AI key or run a local model.
Why I built it
Most system design practice tools put the feedback loop behind an expensive subscription. The underlying interview framework is public knowledge, but students still need structured questions, stage-by-stage guidance, and fast critique. DesignDojo keeps that loop free by moving the model choice and usage cost to the user.
What's inside
- System design and LLD runners with stage definitions, prompts, and rubric-based feedback.
- BYOK model support for OpenAI-compatible providers, Anthropic, Gemini, OpenRouter, Groq, Together, Ollama, LM Studio, and vLLM.
- Excalidraw whiteboard for diagramming during practice sessions.
- Local-first session storage with IndexedDB, so practice state stays in the browser.
- Zod-validated content and AI output for question schemas, framework definitions, and structured feedback.