Custom NLP Model Development

I need an experienced AI developer to build a custom NLP model that answers fact-based questions using data from structured MSSQL tables.

Requirements:
- The NLP model should interpret and respond to fact-based questions only.
- The data source will be structured tables from an MSSQL database.
- Expected output from the model is in the form of textual answers.

Scope of Work:
- Train a lightweight NLP model (e.g., T5-small or similar) to process natural language questions.
- Fine-tune the model on a custom Q&A dataset (we can help with domain-specific examples).
- Set up a Python API (FastAPI preferred) to serve the model.
- Integrate with MSSQL to enable data-driven answers (e.g., sales totals, last month performance).
- Include forecasting logic (using Prophet or similar) for answering forecast-related queries (e.g., “Expected sales next month”).

Deliverables:
- Trained and tested NLP model.
- FastAPI (or similar) service to host the model.
- Forecasting module (Python-based) integrated into the service.
- API documentation to consume the model.
- Basic instructions to retrain or expand dataset in the future.

Ideal Skills and Experience:
- Expertise in NLP and AI model development.
- Proficiency in working with MSSQL databases and structured data.
- Strong background in generating accurate textual responses from AI models.
- Prior experience in similar NLP projects is a plus.