I built an AI tool to practice system design like LeetCode — would love feedback!

#ai #distributedsystems #interview

Hey everyone 👋

I’m a software engineer currently at Google, and for the past few months I’ve been working on a side project called Desyra — a platform where you can practice system design problems interactively, a bit like how you’d solve algorithmic problems on LeetCode.

Instead of just reading solutions, you actually design systems using a drag-and-drop UML canvas, and Desyra’s AI gives you feedback across scalability, reliability, and security — almost like a mentor reviewing your design.

💡 Why I built this
When I was prepping for system design interviews (and later mentoring others), I noticed how hard it is to get feedback.
You can watch YouTube breakdowns, read blog posts, or use diagrams.net — but nothing tells you “how good was your design?”

Desyra tries to fix that by giving you a space to:

Practice system design with real-world prompts

Build and visually diagram architectures

Get automated feedback on your design quality

And eventually, compare your design approach to community solutions

🚀 Current stage
82 problems live (20 are free to try)

Hosted on desyra.io

Still solo — bootstrapping and gathering feedback before I scale further

🙏 What I’d love feedback on
Does this sound useful to you (or your peers)?

Would you prefer a free tier forever or a credit-based model?

Any ideas on how I can make the AI feedback more meaningful or personalized?

I’m staying anonymous for now but happy to share details about the tech stack, the AI scoring approach, or even the mistakes I made along the way if that’s useful for others here.

Would love to hear your thoughts — both as builders and as potential users.

Thanks 🙌

— A solo dev who can’t stop building
desyra.io