Research

Last update: October 2026

My work connects artificial intelligence (AI), software engineering (SE), and computer‑supported cooperative work (CSCW).

Current research themes

01

Human-AI collaboration in software engineering

Effective AI tools should strengthen human judgment, learning, and collaboration, not merely automate tasks.

Selected Findings
  • Output quality, practical value, and goal alignment are associated with developers’ trust in generative AI [TOSEM’26].
  • Project-grounded agents nearly match humans at answering developer questions, but need briefer responses [ICSE’26].
  • Developer-AI tools need varied interaction styles and support for user control [FORGE’25, ICSE’25].
  • Workflow-integrated conversational support helps ML end-user programmers [ICSE’24].
  • Code-review bots increase merges but reduce discussion; automation creates tradeoffs [EMSE’23, EMSE’22].
  • Chatbot language and social characteristics shape user experience [TOCHI, IJHCI].
02

Supporting open source software (OSS) communities

Sustainable OSS communities depend on participation, supportive social practices, and a genuine sense of belonging.

Selected Findings
  • Identical OSS role titles can hide different duties; explicit responsibilities clarify governance [CHASE’26].
  • Mentors, clear onboarding, and suitable tasks support newcomers’ participation [CHASE’25, IST’24].
  • Contributor retention requires timely, context-specific support [TOSEM’25, EMSE’22].
  • Interpersonal barriers undermine belonging, especially for marginalized contributors [ICSE’25].
  • Social connection and intrinsic motivation strengthen belonging in OSS [ICSE’23].
More findings on OSS communities
03

Technology-mediated learning

Effective computing education combines evidence-based pedagogy, authentic software projects, structured guidance, and deliberate support for learning with AI.

Selected Findings
  • Novices treated GenAI’s responses as fixed, rarely refining prompts when explanations fell short [CUI’25].
  • GenAI supports learning, but guidance is needed to address overreliance [CSEE&T’25, ICSE’24].
  • Gamified scaffolding helps students engage with OSS contributions [CSEE&T’25, VL/HCC’24].
  • Authentic OSS projects develop skills but need structured guidance [FSE’25, CSEE&T’17].
  • Instructors want AI tutors that adapt step-by-step guidance to students’ needs [CONVERSATIONS’23].
  • Programming instructors need pedagogical knowledge beyond technical content [ITiCSE’19, FIE’19].
  • Students and instructors perceive programming challenges differently [IEEE LATAM’17].

Research approach

  • Pragmatism Methods tailored to each problem and context, combining quantitative and qualitative approaches when useful.
  • Engineering Real-world problems addressed through purpose-built tools and applications.
  • Innovation New ideas and alternative ways of framing problems.
  • Quality Rigor and quality throughout the research process.
  • Interdisciplinarity Knowledge across domains connected to uncover new problems and opportunities.

Earlier research projects

View selected publications