AI-knows.me

personalized question
reponsive and answers from the backend
google cloud, backend
vercel deployment

🚀 Inspiration

I envisioned an AI-powered platform that truly gets to know you. Our mission is to build a tool that learns about your unique strengths over time, using data from your LinkedIn profile and, in the future, from sources like your CV, Instagram profile, and even public online presence (e.g., Google search results). This way, AI-knows.me can deliver personalized insights that help you excel in every aspect of your life.

🔍 What It Does

AI-knows.me is a dynamic, AI-driven platform that: • Uncovers Your Strengths: Uses an adaptive Q&A process to reveal your unique talents and interests. • Personalized Insights: Generates recommendations tailored to your career, hobbies, lifestyle, and startup ideas. • Data Integration: Initially pulls basic profile data from LinkedIn (and, eventually, will expand to include your CV, IG profile, and more). • Mission: As the AI learns more about you, it will offer increasingly nuanced insights, helping you make informed decisions about your professional and personal growth.

🛠️ How I Built It

And no, its not a chatgpt wrapper! • Frontend: Built with Next.js (TypeScript) for a modern, responsive, and interactive user experience using pure css to have 0 problems with UI. • Backend: Developed using FastAPI and containerized with Docker, then deployed on Google Cloud Run. • Authentication: Integrated LinkedIn authentication via NextAuth, ensuring secure sign-ins. • Adaptive AI: Implemented a custom rule-based Q&A flow with sentiment and keyword analysis to generate tailored recommendations. • Deployment: Frontend is hosted on Vercel and backend on Cloud Run, providing scalability and reliability going through github first and then vercel so every update is live instantly to the domain name ai-knows.me (used hostinger).

⚙️ Challenges I Ran Into

• Deployment Configuration: Fine-tuning Docker and Cloud Run settings (e.g., memory limits and port configuration) to be able to publish.
• OAuth Integration: Ensuring the LinkedIn OAuth callback URL matched exactly with the one registered in the LinkedIn Developer Portal idk why this is a problem tho I think OAuth has some error there.
• Adaptive Flow Complexity: Developing and debugging the dynamic Q&A system to accurately reflect and analyze user responses, its kinda hard doing dynamic Q with a simple AI model used here.
• Data Integration: Securing and seamlessly integrating various data sources to produce reliable insights and getting the requests to work.

🏆 Accomplishments I'm Proud Of

• Successfully building and deploying a full-stack, AI-powered platform without relying on heavy ChatGPT wrappers.
• Creating a robust, adaptive Q&A system that learns from user responses.
• Seamlessly integrating LinkedIn authentication for a secure user experience.
• Establishing a scalable deployment pipeline using Docker, Google Cloud Run, and Vercel.
    • Doing this as a CS freshman and alone (19 y old) was truly a challenge

📚 What I Learned

• The critical importance of proper configuration in production environments.
• How to integrate lightweight AI models for real-time sentiment and keyword analysis.
• Best practices for creating adaptive and personalized user experiences.
• Strategies for securely managing third-party OAuth integrations.

🔮 What’s Next for AI-knows.me

• Enhanced Data Integration:

In the future, I plan to incorporate more data sources such as your CV, Instagram profile, and even public online data to generate even richer insights. • Advanced AI Models: I aim to integrate more sophisticated AI models that provide nuanced, detailed recommendations. • Expanded Personalization: Expect future versions to offer insights across additional areas like personal branding, networking, and learning opportunities. • Improved UI/UX: I'll continue refining the user interface based on feedback, ensuring a seamless and engaging experience for every user. • Scalability & Performance: I’ll optimize our infrastructure to handle a larger user base and more complex data integrations.