The core of AI4SE consists of five research tracks in which researchers and five PhD students (one for each track) work together on a range of relevant topics. The research partnership will be run by a joint team of researchers at JetBrains and TU Delft, including Prof. Dr. Arie van Deursen and Dr. Maliheh Izadi from TU Delft, as well as Dr. Vladimir Kovalenko and Dr. Pouria Derakhshanfar from JetBrains.
“As a software engineering researcher and educator, I am thrilled about the new partnership between TU Delft and JetBrains. The software development tools produced by JetBrains are loved and used by software engineers across the globe. Within the AI4SE lab, TU Delft and JetBrains will join forces to take full advantage of the latest developments in AI to revolutionize software development.”
“A core value of JetBrains is to provide efficient solutions for complex tasks and problems through cutting-edge technology. The new AI4SE partnership allows us to join forces with top scientists to make the process of creating software an even more enjoyable, creative, and productive experience. By integrating our expertise in building top-notch software engineering tools with our practical applications of AI and TU Delft’s advanced AI research capabilities, we aim to achieve unparalleled results.”
“AI4SE is a big new chapter for us at JetBrains Research. We are excited to be working shoulder to shoulder with the talented scientists and students of TU Delft. We are looking forward to making AI4SE a fruitful collaboration with long-term impact far beyond JetBrains and the research community.”
If you are interested in joining our research team as a Ph.D. student, you can apply via the links to each vacancy included in the track descriptions.
This track focuses on code generation using AI techniques. We aim to explore the strategies for validating this generated code and how this generated code can help developers validate their handwritten code.
This track aims to refine generic large language models for code to suit various scenarios. By tailoring these models to the specific needs of individual users, projects, and organizations, we can ensure personalized outputs. The models will be optimized to produce legal, safe, and timely predictions and operate efficiently on low-resource devices.
This track aims to embed emerging large language model practices, such as code generation or code explanation, into developers' workflows without disturbing users to improve their productivity. To do so, we will study user interaction with the model in IDEs, research the user experience, and investigate how to best build trust between developers and their intelligent agents.
This track aims to seamlessly integrate runtime information into JetBrains IDEs, elevating the development experience by enhancing code quality, pinpointing and addressing performance issues, and providing precise code assistance within the IDE environment. To achieve this, we will bridge the gap between static and dynamic information within machine learning techniques.
This track aims to develop an intelligent, AI-based student assistant that provides context-informed support for programming education while stimulating knowledge transfer. We will focus specifically on how automatically generated hints can guide students toward the right solution, how students interact with the intelligent assistant, and how efficient their interactions are toward their learning objectives.