AI-Driven Natural Language Query Suggestions

• Developed Natural Language Query Generator, to help users get meaningful insights from their data through natural language search powered by LLMs. Presented to Sr. Director Data Engineering, Hong Wang.

• Optimized legacy code by spearheading the use of Retrieval Augumented Generation and Neo4j to deliver context-aware zephyr-7b, and reduced 2000+ lines of code to less than 600.

• Achieved an average Kullback-Leibler divergence score ranging from 9.3 to 11.67, indicating high relevancy of generated prompts.