Research Companion.
A platform to help faculty organize and interact with research papers, developed in collaboration with Mridul Joshi, Researcher at Stanford.
More time for the research that matters.
Illustrative workflow · explore the engineering above
The problem
Organizing a research library required manual metadata entry, while finding relevant papers meant repeatedly searching through documents. The workflow needed to make a growing collection easier to maintain and query.
What I built
Built the RAG service with Python and FastAPI so researchers could query document libraries with context-aware responses. Vector-based search retrieved papers by conceptual similarity. A MongoDB, Express, React, and Node.js backbone handled automated metadata extraction and library organization.
The engineering thinking
Conceptual similarity makes relevant papers discoverable even when the question and document use different words. Automated metadata extraction supports that experience by reducing the manual work needed to maintain the library.
The outcome
Streamlined the research workflow and made information retrieval more intuitive and context-aware. The metadata pipeline was adopted as the group’s standard literature-review workflow.