← Work

2025 · AI assistant

A natural-language SQL assistant for a data analyst

a mobile services company · Ai/ML engineer

The problem

Data analysts needed answers from several databases without always knowing the table and column names.

What I built

I built an assistant that reads each database's structure first, then asks a language model to write the query with that context. It runs the query and shows the result in a simple web app.

The result

Analysts stopped looking up table and column names before every question.

Main tools

  • Amazon Bedrock
  • Claude
  • SQLAlchemy
  • Streamlit
  • Python

This is the short version. The full case study is available on request.

Ask for the full case study →