Accuracy
Context-aware SQL generation
Improved SQL generation accuracy for context-driven tasks by refining prompts and agent execution flow.
- Role
- AI Engineering Intern
- Duration
- May 2024 - July 2024
- Location
- Hyderabad, On-Site
- Focus
- Context-aware SQL generation, chain-of-thought prompting, ReAct agents, and LLM benchmarking.
Reasoning
Chain-of-thought + ReAct
Applied chain-of-thought reasoning and ReAct agent strategies to increase reliability across query scenarios.
Evaluation
Custom LLM benchmark framework
Designed a benchmark framework to evaluate, compare, and rank models on context-focused SQL workloads.
Outcome
Better model and prompt decisions
Created a repeatable evaluation process for selecting model and prompting combinations based on evidence.
Certificate and internship documents
Embedded internship document: