AI NLP Model with SQL Integration

I need a robust Natural Language Processing solution that talks directly to my existing SQL database. The idea is to create, train, and deploy an NLP model whose primary task is database integration—pulling raw text from specified tables, processing it, then writing structured results back into new or existing fields so downstream analytics can use them instantly.

Here’s the flow I have in mind:
• Text records are fetched from the SQL layer.
• The NLP engine processes each record—tokenising, cleaning, running the core model—and packages the output in an easy-to-query format.
• Results and any relevant confidence scores are inserted or updated in the same database with minimal latency.

You’re free to choose the framework—Python with spaCy or Hugging Face Transformers makes sense, though if you prefer TensorFlow or PyTorch that’s fine as long as setup steps are documented. .

Deliverables are straightforward:
1. Well-commented source code for the model and integration layer
2. A requirements.txt or environment.yml for reproducibility
3. One-click run or deploy instructions and a short README
4. A sample SQL script that creates the target tables and demonstrates end-to-end functionality

I’ll validate the work by running the script on a staging database; the task is considered complete once processed records appear with the expected output and no data loss.