Automated QA suite for enterprise Sales Data Agents. Automates ground truth generation and AI-driven answer evaluation to ensure NL2SQL accuracy.

A controllable NL2SQL / SQL agent project with CLI, API, guard rails, and evaluation workflow.

agent for nl2sql with capabilities for analysis and database lookups, visualization, multi-hop reasoning and evaluation

Agentic LLM-based NL2SQL system with multilingual & voice input, schema-aware retrieval, validation, and Spider benchmark evaluation.

The first execution-oriented benchmark for heterogeneous data retrieval (NL2SQL/API/Code). Evaluating LLMs & Data Agents based on real-world business results, not just syntax.

A modular NL2SQL agent implementing Chain-of-Thought reasoning and schema refinement, evaluated on the BIRD-Bench dataset. Inspired by SQL-R1 and Long-Context LLMs.

End-to-end AI NL2SQL system using Vanna 2.0, FastAPI, and SQLite. Translates natural language into SQL with validation, executes queries, and visualizes results using Plotly. Includes agent memory, edge-case handling, and evaluation on 20 test queries.

End-to-end AI NL2SQL system using Vanna 2.0, FastAPI, and SQLite. Translates natural language into SQL with validation, executes queries, and visualizes results using Plotly. Includes agent memory, edge-case handling, and evaluation on 20 test queries.