-
-
Generated Website Using Image and Resume
Inspiration
While LinkedIn profiles are the standard nowadays, personal websites are exploding as a means of marketing oneself to employers while standing out in a unique and versatile way. By combining of LinkedIn, Wix, and AI-powered LLMs, we can create AI website builders that can create custom websites, specifically curated for potential employers.
What it does
Our platform generates a personalized website using AI. Users can upload a profile picture and their resume in PDF format. After uploading, they can click the "create" button to see a preview of their personalized website. Once satisfied with the design, users can export their site to HTML, allowing them to download their website as an HTML file. Currently, our submission has a simple mock template as a demo proof of concept, but we are working on more custom templates and editable features to make it as versatile and personalized as possible.
How we built it
We used Flask to handle backend operations and the Google Generative AI API to read and analyze a resume or any other information source from the pdf. After analyzing, Google Gemini generates a consistently structured response, which we parse for patterns and extract data from, passing on a jsonified Python dictionary to the frontend. The frontend is built with HTML, CSS, and JavaScript, utilizing Bootstrap for styling.
Challenges we ran into
We faced difficulties in tuning parameter values to get consistent responses from Gemini AI and encountered issues with making the website downloadable in HTML. Exporting as HTML is also bad for scalability, if we want users to have custom website templates, which led us to pivot towards using a subdomain system for easier implementation.
Accomplishments that we're proud of
We successfully built a functional website that integrates AI to achieve our goal. We managed to parse resume content effectively and implement dynamic profile generation, which reads PDF resumes, extracts relevant information, and generates personalized HTML profiles. This highlights our capability to work with APIs, handle file uploads, and dynamically manipulate DOM elements.
What we learned
We deepened our understanding of the Flask framework and improved our skills in JavaScript event handling. We also learned about working with Blueprints in Flask and how client-side JavaScript communicates with server-side Python scripts through HTTP requests and JSON data exchange.
What's next for PortfolioPro
We plan to add customizable templates with drag-and-drop editors and other interactive elements. Additionally, we aim to implement a user account system to allow users to save their websites. Leveraging AI for content suggestions and additions, as well as incorporating a rating system for resumes to help recruiters, is also on our roadmap. These enhancements will make the platform even more valuable for both job seekers and employers.
Log in or sign up for Devpost to join the conversation.