I am an applied scientist at Salesforce Research working on conversational AI, with a focus on evaluation of LLMs and LLM-based agents.
Some highlights from my previous work:
- Demonstrate the universality of sycophantic behavior in LLMs and provide a robust framework to evaluate solutions.
- Provide Conversational Recommender Systems with access to external knowledge to fullfill an educational objective in addition to recommendations.
- Build a pair of LLM agents to simulate seller and shopper interactions in the e-commerce setting.
- Explore a solution to accompany news readers in discovering coverage diversity.
- Trained state-of-the-art question generator by diversifying the types of answers utilized.