Custom AI-Powered Web App

I am building a cloud-hosted web application whose primary feature is a bespoke text-generation engine rather than an off-the-shelf GPT-3 or GPT-4. The customised model must run inside Azure, draw fresh context from my own data through a robust RAG (Retrieval-Augmented Generation) pipeline, and surface the results in a modern single-page app that also embeds a Power BI custom visual.

The work involves four tightly-connected layers:

• Model layer – fine-tune or train the customised LLM, containerise it, and expose it through Azure OpenAI / Azure ML for low-latency text generation.
• Retrieval layer – hook the model to Azure Cognitive Search (or a comparable Azure service) so user prompts are enriched with content pulled from our SQL database and SharePoint document store.
• Application layer – build the front-end (React or Blazor is fine) that captures prompts, streams responses, and presents supporting citations.
• Insight layer – craft a Power BI custom visual that sits inside the same portal and plots usage analytics in real time.

Acceptance criteria
• A running Azure deployment accessible over HTTPS with AAD-protected login.
• REST/GraphQL endpoints that return model output within the agreed latency budget.
• RAG pipeline demonstrably injecting up-to-date records from the database.
• Power BI visual packaged and published to my workspace, reading from the app’s telemetry table.
• Code, IaC templates, and a concise hand-over guide committed to my private Git repo.

If you are fluent in Azure AI, LLM fine-tuning, modern JavaScript/C#, and Power BI custom visuals, Python, fastapi this project should feel familiar. I am ready to start as soon as we agree on a delivery schedule.