Research collaboration is crucial to the development of researchers’ careers as well as building the capability to solve larger problems and accelerate scientific discovery. Researchers and research institutions are busy and often isolated from other researchers even within similar fields. To address this challenge, the author was tasked with identifying and recommending collaboration opportunities in a fast, easy-to-use, and platform-agnostic interface. Leveraging new large language models trained specifically on research papers, the author created a corpus of encoded 2022-2023 NSF-funded research projects and integrated it into a web-based application. The resulting application enables users to input any given research title and abstract and quickly identify similar projects from the encoded corpus. The application also provides other useful information about the matching results such as principal investigators, funding programs, and a map of institutions doing similar work. In this paper, the author discusses the development of this valuable application and why modern LLMs are well suited to this task. The author also highlights potential areas for future expansion and development. This including expanding the corpus to more years of funded projects and more agencies’ funded projects, accepting user feedback on result relevance, and identifying collaborations earlier in the grant proposal submission process.