Nihal Prabhu

Data Science Intern @UsefulBI Corporation

San Ramon, CA, US

Optimized AI LangChain agent loop execution by reducing redundant calls and improving prompt construction, lowering agent inference latency by ~100ms per task cycle, and reducing token usage.Deployed scalable generative AI agents by integrating LangChain with AWS Bedrock Converse and EC2, enabling low-latency, real-time interactions.Developed RAG system connected to a vector database, enabling it to utilize the most relevant information when answering questions, improving response accuracy, and reducing off-topic answers.Evaluated AI workflows using scikit-learn to score agent responses based on retrieval relevance, precision, recall, and F1 metrics to benchmark grounding consistency.

University of California, Davis

Bachelor of Science - BS