Graduate Researcher – Search and Ranking

Evaluating Retrieval Methodologies to Augment LLM Event Forecasting

Description

I worked at the Agentic Lab under advisement from Prof. Mengye Ren wherein I was responsible for the developing the system design and infrastructure for facilitating fast experimentation with various RAG pipelines as well as exploring and evaluating retrieval performance for LLM forecasting performance over daily news data.

Stack: PyTorch, vLLM, Hugging Face, SentenceTransformers, Milvus, Singularity, SLURM, Git