When I applied for a recent internship, I didn’t expect the assignment they gave me to be a whole project in itself.
I was asked to build a full-stack blog website, but here’s the catch — it had to be built with MongoDB, Mongoose, and Next.js API Routes. Up until that point, I was much more comfortable using Prisma, PostgreSQL, and Server Actions.
But I took the challenge head-on — and I’m glad I did.
🚀 What I Built
I ended up creating the Truly IAS Blog — a modern, clean, and content-rich blog platform where:
- ✍️ Rich-text editing is powered by Tiptap
- 🖼️ Blog images are uploaded via Cloudinary and automatically content-aware cropped for thumbnails
- 🔐 Auth is handled by Clerk
- 🌓 Dark and light mode UI built with Shadcn/UI and Tailwind CSS
You can check out the Live Demo or Source Code.
While I wasn’t selected for the internship, I walked away with something far more valuable — experience. It forced me out of my comfort zone and helped me understand MongoDB and Mongoose deeply.
🧠 What I Plan to Do Next
This project won’t just sit in my GitHub, because v2 is coming. And here’s what I’m planning:
- Migrate to Prisma + PostgreSQL
- Add Server Actions for more seamless backend logic
- Expand the schema with categories, authors, trending posts
- Improve the UI
- Add AI assistance for content creation (more on that in a second!)
- Integrate admin panel analytics with charts
🎯 My AI Learning Plan
For a long time, I wanted to integrate AI into my projects — but there was one major blocker: OpenAI credits.
Well, not anymore. I finally bought some, and now I’m diving into the world of AI integration with full focus.
I already have an idea I’m working on — a SaaS tool where users can:
🔗 Input a YouTube video link or upload a video file →
🧠 Get an SEO-optimized blog post based on the content
Here’s the current progress:
- ✅ Using Whisper-1 to transcribe videos (and yes, the accuracy is surprisingly good!)
- 🔜 Next steps: Feed that transcript into a powerful LLM for blog post generation
This project is going to teach me how to use AI in production — and I’m super excited to integrate similar functionality in Truly IAS Blog, Dropify, and Invoicepedia once my exams wrap up at the end of July.
🔄 Learning Background Jobs & Queues
One thing I’ve realized — when building AI SaaS tools like this, you can't keep the user waiting on a loading screen forever.
That’s why my next big learning goal is understanding background jobs and queuing systems using tools like Inngest.
I want to figure out how to:
- Schedule and run background jobs efficiently
- Queue transcription + blog generation flows
- Show real-time progress/status to users
🧱 What’s Driving Me
Right now, my only focus is on learning as much as I can by building things.
I’ve got 3 months post-exams, and I’m dedicating that time fully to:
- Deepening my backend knowledge
- Mastering AI integrations
- Understanding queuing and job scheduling
- Creating real products that solve real problems
If you’re also in this learning-by-building phase, let’s connect — or just drop a message. Would love to exchange ideas or collaborate on projects.
More updates coming soon 🚀
— Sidharth Sangelia