Hi,

I found out about this task just 4 days ago, but because of my job, I couldn’t start right away. I created the GitHub repo and basic setup that day, but I only started building properly on Sunday.

I wanted everything to be perfect, so I went a bit overboard. You said we could use mock data — but I built a real scraper using Playwright to get live credit card data from bank websites. You asked for GenAI-based chat — I added more features like card comparison, eligibility check, and smart search using Gemini AI.

Because of all that, the project is still not fully finished. I’m submitting it now to meet the deadline, but I’ll keep working on it. If you check again on Wednesday, it’ll be more polished. Right now the code is a bit messy because things got crazy while building.

Also, I did use AI tools like ChatGPT to help with some UI parts and small helper functions — not the whole thing, just to speed things up.

👨‍💻 About Me

I’m Pranav Bhatkar, a self-taught full-stack developer from Maharashtra.

I started as an office boy, but now I’m a developer at PPS Energy Solutions. I built their full office management system.

I’m also worked at Axentia as a developer intern. I’ve built AI-based tools there like:

CollegeGenie – a chatbot for college help (Telegram + WhatsApp bots using OpenAI)

Super Teacher AI – for teachers to generate lesson plans, report cards, and more using AI

I handle both frontend (Next.js) and backend (NestJS, Postgres).

I also build mobile apps, do freelance work, and enjoy solving real-world problems with code.

💳 About CardSift

🔗 Live App | GitHub

Built using Next.js 15, Tailwind CSS, and BetterAuth

Uses Playwright to scrape real credit card data from banks

Uses Gemini AI to answer user questions like:

"Best cards for fuel cashback?"

"Compare HDFC Regalia and Axis Magnus"

Stores data in PostgreSQL

Includes user login and a simple eligibility checker

🚧 What’s Next

Clean up messy code

Improve UI

Add offer history for each card

Add smarter card suggestions using AI

Thanks for checking out my work. I only had 1.5 days to build this, but I gave it everything I could. I’ll keep improving it even after submission.

Pranav