Speaker "Divyansh Agarwal" Details Back



Building and evaluating Generative AI applications for NLP


The talk will focus on the different aspects of building interactive AI systems that use large language models (LLMs). We will look into the trend of LLM development and the evaluation of systems from the lens of academic research in NLP. The talk will go over how effective LLMs like chatGPT, Claude, Bard etc. are, at different problems in natural language generation space. We'll go over how generative AI has brought focus back to the data and ethical development of AI Systems. We will discuss the field of HCI-NLP, and the successful design of Generative AI applications.


Divyansh is a Sr. Research Engineer with the conversational AI team at Salesforce Research. His work involves developing and studying AI applications that allow real-world user interaction with LLMs. He is active in NLP research on topics like text summarization, factuality and natural language generation. Divyansh graduated from the Language Technologies Institute (LTI) at Carnegie Mellon University (CMU) in Pittsburgh. Divyansh is interested in the interaction of humans with AI technologies, the interplay of ethics and privacy in the mix, and how that can inform the future generation of AI applications.