As the lead developer of Sketch to SaaS, I was responsible for the end-to-end development of the platform, focusing on the bridge between AI vision and frontend execution. My key contributions include:
AI Pipeline Integration: I architected the connection with the Google Gemini 3.0 (2.0 Flash) API, implementing the logic to convert raw image data into Base64 strings for real-time multimodal processing.
System Prompt Engineering: I designed and fine-tuned the complex System Instructions that guide the AI to generate clean, production-ready Tailwind CSS code while strictly avoiding conversational filler or formatting errors.
Dynamic UI/UX Development: I built the entire responsive frontend using HTML5 and Tailwind CSS, implementing a premium Glassmorphism design language and custom CSS animations for a high-end SaaS feel.
Sandbox Rendering Logic: I developed the secure Iframe Sandbox system that allows the AI-generated code to be injected and rendered instantly in a "live browser" environment without reloading the main page.
Simulated Auth System: I created the interactive Login/Register modal and logic that simulates a real user onboarding experience, including dynamic Navbar updates to enhance the project’s "MVP" (Minimum Viable Product) feel.
Error Handling & Optimization: I optimized the generation workflow to achieve sub-10-second response times and implemented error-catching to handle API latencies or invalid image uploads.
Log in or sign up for Devpost to join the conversation.