Develop a sophisticated, modular AI system capable of integrating and orchestrating multiple state-of-the-art AI models. The system should be designed to accommodate a wide range of model types, including but not limited to, large language models (LLMs), image generation models, audio processing models, and specialized models for tasks like code generation or scientific simulations. User-provided API keys should be the primary method for integrating these models, ensuring flexibility and user control. Prioritize a design that allows for easy addition and removal of models, enabling the system to evolve as new technologies emerge.

At the core of this system, create a 'Master AI' that functions as a root controller. This Master AI should possess the following capabilities:

1. Task Decomposition and Routing: The Master AI must be able to analyze user requests, break them down into sub-tasks, and intelligently route these sub-tasks to the most appropriate integrated models. It should consider the strengths and limitations of each model, optimizing for accuracy and efficiency.
2. Data Orchestration and Fusion: The Master AI should manage the flow of data between models, ensuring seamless communication and data transformation. It should also be able to fuse outputs from multiple models to generate a coherent and comprehensive response.
3. Contextual Awareness and Memory: Implement a robust memory mechanism that allows the Master AI to maintain context across multiple interactions and tasks. This memory should be accessible to all integrated models, enabling them to leverage past information for improved performance.
4. Error Handling and Fallback Mechanisms: Design the system to handle errors gracefully, implementing fallback mechanisms to ensure that user requests are processed even in the event of model failures or API issues. The Master AI should be able to identify and report errors to the user, providing clear and actionable feedback.
5. User Interface and API: Provide a user-friendly interface for managing integrated models and interacting with the Master AI. Additionally, offer a comprehensive API for programmatic access, enabling developers to integrate the system into their applications.
6. Security and Privacy: Prioritize security and privacy by implementing robust authentication and authorization mechanisms. Ensure that user API keys and data are stored securely and that the system adheres to relevant privacy regulations.
7. Model Selection and Optimization: Develop a system where the Master AI can optimize the model selection process based on the user's specific request. Allow the user to also be able to set preferences for model selection, and prioritize models based on cost, speed, or accuracy.

The system should be designed for scalability and maintainability, allowing for future expansion and updates. Thoroughly document the system's architecture, APIs, and usage instructions to facilitate user adoption and development."

Currently use in backed py Only give me this py code as you detailed fetures and etc, I hope you understand. Current I ittriagate DeepSeek and Gemini, this is my master ai, i use this both ai api and after get Master converted best response get. I helped you understand as get fetures all add in my this master ai