BRIDGING DESIGN AND DEVELOPMENT: BUILDING A GENERATIVE AI PLATFORM FOR AUTOMATED CODE GENERATION
DOI:
https://doi.org/10.34218/IJCET_16_02_038Keywords:
Generative AI, Code Generation, Large Language Models (LLMs), Low-Code Development, Prompt Engineering, Software Automation, AI-Assisted Development, Design-to-Code, Intelligent IDEs, Software Engineering ProductivityAbstract
The disconnect between design and development phases in software engineering often leads to increased development cycles, misinterpretations, and reduced productivity. Recent advancements in Generative AI, particularly large language models (LLMs), offer promising capabilities for automating code generation directly from high-level design artifacts or natural language prompts. This paper presents the architecture and implementation of a generative AI-powered platform designed to bridge the gap between UI/UX design and functional code development. The platform integrates components such as prompt engineering layers, pre-trained LLMs, design parsers, and code validators to convert design inputs into production-ready code. We evaluate the system using two real-world use cases: automatic transformation of web form designs into ReactJS code and full-stack application scaffolding from Figma prototypes. Our experimental results demonstrate significant reductions in development time and manual effort, with an average code generation accuracy exceeding 85%. Additionally, the platform enhances collaboration between designers and developers by streamlining the transition from mockups to executable components. The findings highlight the potential of generative AI in accelerating software delivery, reducing human error, and enabling rapid prototyping in modern development environments. Future enhancements include support for multi-modal inputs, continuous learning, and integration with CI/CD pipelines.
References
Chen, M., et al. “Evaluating Large Language Models Trained on Code.” arXiv preprint arXiv:2107.03374, 2021.
OpenAI. “Introducing GitHub Copilot.” https://github.com/features/copilot
Li, Y., et al. “CodeT5: Identifier-aware Unified Pre-trained Encoder-Decoder Models for Code Understanding and Generation.” arXiv preprint arXiv:2109.00859, 2021.
Touvron, H., et al. “Code LLaMA: Open Foundation Models for Code.” Meta AI, 2023.
Svyatkovskiy, A., et al. “Intellicode Compose: Code Generation Using Transformer.” arXiv preprint arXiv:2005.08025, 2020.
Jain, A., et al. “Efficient Design-to-Code Generation for Responsive UIs.” Proceedings of the CHI Conference on Human Factors in Computing Systems, 2022.
Weng, L., “Prompt Engineering Techniques.” Lil’Log, 2023.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Chandra Shekar Chennamsetty (Author)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
