Register now After registration you will be able to apply for this opportunity online.
GenAI supported code analysis for automated UI generation
Automated UI generation through Generative AI (GenAI) technology offers an interesting approach to designing user interfaces directly from the source code with minimal human input. The hypothesis is that GenAI can improve the analysis of user requirements, design patterns, and usability principles to autonomously create UI components and layouts. As a result, developers can create user interfaces much faster, ultimately supporting rapid prototyping, and in doing so, improving overall product quality.
Keywords: UX/UI, UI, GUI, Automation, GenAI, Generative AI, LLM, Large Language Models
In traditional industrial sectors such as mechanical engineering, a lot of valuable code is created but neither shared nor tested simply because of code deployment being too complex or tedious. There-fore, further automation of any process related to code deployment – such as UI generation – is de-sired. The recent progress in the field of GenAI may support this development and, thus, should be further explored within this project.
First of all, a thorough analysis of existing tools for GenAI supported code parsing and automated UI generation needs to be conducted. Based on that, the code analysis capabilities of “non-AI” and “AI supported” tools shall be evaluated and areas for meaningful GenAI application shall be identified. Provided that the use of GenAI adds value to the process of UI generation, a corresponding concept shall be defined – ideally resulting in a first prototype of a GenAI agent.
In addition to the technical aspects, this project offers a unique opportunity to gain insight into the workings of a young software start-up, giving you exposure to the fast-paced environment of an ETH Spin-Off.
In traditional industrial sectors such as mechanical engineering, a lot of valuable code is created but neither shared nor tested simply because of code deployment being too complex or tedious. There-fore, further automation of any process related to code deployment – such as UI generation – is de-sired. The recent progress in the field of GenAI may support this development and, thus, should be further explored within this project. First of all, a thorough analysis of existing tools for GenAI supported code parsing and automated UI generation needs to be conducted. Based on that, the code analysis capabilities of “non-AI” and “AI supported” tools shall be evaluated and areas for meaningful GenAI application shall be identified. Provided that the use of GenAI adds value to the process of UI generation, a corresponding concept shall be defined – ideally resulting in a first prototype of a GenAI agent. In addition to the technical aspects, this project offers a unique opportunity to gain insight into the workings of a young software start-up, giving you exposure to the fast-paced environment of an ETH Spin-Off.
•Screening of available tools for GenAI supported code parsing, automated UI generation, and for related GenAI agents
•Evaluation of code analysis possibilities (non-AI vs. AI-supported) with respect to all areas that are relevant for automated UI generation (e.g. syntax or context-aware parsing, pseudo code, data/control flows, required inputs)
•Define a concept from code analysis to UI generation
•Create a prototype of a GenAI agent for automated UI generation (ideally containing code explainer/docstring generation)
•Screening of available tools for GenAI supported code parsing, automated UI generation, and for related GenAI agents •Evaluation of code analysis possibilities (non-AI vs. AI-supported) with respect to all areas that are relevant for automated UI generation (e.g. syntax or context-aware parsing, pseudo code, data/control flows, required inputs) •Define a concept from code analysis to UI generation •Create a prototype of a GenAI agent for automated UI generation (ideally containing code explainer/docstring generation)
20250208_StudentProject_MetaCode.pdf
