Chatbot refactoring from OpenAI to Google Gemini AI Description: We are looking for an experienced developer to perform a crucial technical refactoring on an existing chatbot (v1.7). The main objective is to migrate the AI engine from OpenAI to Google Gemini and establish fully dynamic management of the bot's configuration and knowledge, using Google Drive and Google Sheets. Required Development Tasks: Migration to Google Gemini 1.5: Replace the OpenAI library with the Google Gemini library, using Vertex AI or the Google AI SDK. This involves adapting the calls and response handling to the new AI engine. Dynamic Model Configuration: The Gemini model ID (e.g., gemini-1.5-flash or gemini-1.5-pro) should not be hardcoded in the backend. It is necessary to implement the ability to read this ID from a configuration variable or, preferably, from a specific cell in the client's control Google Sheet. This will allow the client to change the Gemini model without modifying the code. Self-Managed System Prompt: The bot's initial "system message" or prompt, which defines its personality and rules, should be dynamic. At the start of each session, the bot should consult a specific Google Doc in Google Drive to obtain its instructions. This way, the client can modify the bot's personality or rules simply by editing the document in Drive. Direct RAG from Drive Folders: Leveraging the existing connection with Google Drive, the bot must be able to use files (e.g., PDFs) located in folders designated as knowledge sources for Retrieval Augmented Generation (RAG). If a new document is added to a Drive folder, the bot must immediately integrate it into its context, either by leveraging Gemini's extensive context window or through a simplified search strategy. The project requires a developer with experience in AI API integration, cloud service management (especially Google Cloud Platform), and robust backend development. The ability to implement scalable and easily maintainable solutions will be highly valued. Existing Components (No development required): API connection with Google Drive: The system already has Service Account credentials and can read files from specific folders. API connection with Google Sheets: User verification (Whitelist) and lead registration functionality are operational. Frontend and Security: The React user interface and the OTP (SMTP) authentication system are fully developed and do not require aesthetic or flow modifications. Please provide your Linkedin profile.