Harsh Vassaram
Computer science
Hometown: Maputo, Maputo, Mozambique
Graduation date: Spring 2026
Additional details: Transfer student
FURI | Spring 2026
Evaluation of chunking and storing strategies for increased Retrieval-Augmented Generation (RAG) accuracy
Long context windows in Large Language Models are computationally expensive, and suboptimal context retrieval in Retrieval-Augmented Generation (RAG) produces inaccurate responses or hallucinations. To address this issue, this research proposes a data-backed decision flow for choosing optimal chunking strategies based on the specific domain corpus being ingested for chunking. This approach will evaluate various strategies, to illustrate; the individual functionality, advantages, disadvantages, and limitations. Conclusively, this framework will facilitate engineers to select the most efficient and accurate chunking strategy for their specific RAG implementation, with the aim of maximizing overall performance, whilst minimizing computational resources utilized.
Mentor: Jia Zou
It’s hip to be square.
Students presenting projects at the Fulton Forge Student Research Expo are encouraged to download this personal QR code and include it within your poster. This allows expo attendees to explore more about your project and about you in the future.
Right click the image to save it to your computer.