The goal of Vincent Hellendorn's work is to make safe, effective programming more accessible through intelligent software assistance tools. AI-based tools hold great promise for extracting expertise from public data about software (on GitHub, StackOverflow, etc.) and presenting this in a timely fashion to help new and inexperienced developers.
His research works towards this goal across three frontiers:
- investigating new applications of AI in SE to understand the community’s needs broadly and provide practical value (e.g., FSE’18, FSE’21 IRV, EMSE’21)
- analyzing current models and trends to make sure our results are impactful (ICSE’19, FSE’21, CACM’22)
- building new models based on these insights (ICLR’20, NeurIPS’21, ICLR’22), including PolyCoder, currently the largest open-source model trained exclusively on source code