Following up on the previous article about using Gemini Build to create interfaces for AI-powered systems, I've developed an AI-assisted resume enhancement system. Gemini Build generated approximately 80% of the front-end for this project, and I completed the remaining work by separately developing the backend locally.

Previous Article: Gemini Build Platform Experience

How the System Works:

• Users upload their resume (JPG/PNG image, PDF document, or TXT text file).
• Text is extracted from the resume by Gemini, adhering to a predefined structured JSON schema.
• The resume content is then segmented into editable, deletable, and addable sections and fields.
• Users can currently leverage AI to enhance their career objective and job responsibilities sections.
• Finally, users can select a template and export their resume as a PDF.
• Users can also build a resume from scratch by manually populating its sections.

---

Technologies Used:

• Front-end: Vite, React, React Router, TailwindCSS, hosted on Cloudflare Pages.
• Back-end: Hono, Cloudflare D1, Cloudflare KV, Cloudflare Browser Rendering, with the system hosted on Cloudflare Workers.

The system dynamically switches between three AI models based on load or error conditions, aiming to conserve the daily free usage quota. For instance, it starts with Gemini 2.5 Pro, then shifts to Gemini 2.5 Flash, and finally to Gemma 3.

Additional Notes:

• The system is currently in its initial version, with ongoing work to refine its functionalities.
• The system is freely available to everyone.
• Login is via Google account.
• User resume data is never permanently stored; it is only held temporarily for processing.
• Only the login email address is stored, and it will not be shared with any third party under any circumstances.

Platform Link: https://tahseen.wdalhaj.me