Project: Playground

From Girikon Solutions PVT Ltd (Sep 2023 - Feb 2025) [cite: 201, 204]

Business Problem Statement

Non-technical users faced difficulties accessing and querying various databases due to a lack of SQL syntax knowledge. This created a barrier to effortlessly extracting data insights. [cite: 206]

Solution

• Developed an LLM AI agents model that converts natural language queries into targeted data extraction queries, enhancing user interaction with datasets. [cite: 205]
• Enabled non-technical users to effortlessly access and query various databases using plain language, removing the barrier of SQL syntax knowledge. [cite: 206]
• Facilitated the extraction of data insights from multiple formats, including Excel, Google Sheets, and CSV files. [cite: 207]
• Achieved a query accuracy of 75%, ensuring reliable and relevant data outputs for end-users. [cite: 208]
• Streamlined data retrieval processes, significantly reducing the time required to obtain insights. [cite: 209]
• Improved data accessibility for diverse user groups, fostering a data-driven culture within the organization. [cite: 210]

Tools & Frameworks Used

• Large Language Models (LLMs) [cite: 205]
• Python [cite: 338]
• Pandas, NumPy [cite: 324, 358]
• Scikit-learn, TensorFlow, PyTorch, Keras, Transformers, LangChain, LlamaIndex, LangGraph, LangSmith (general AI/ML libraries) [cite: 351]