Vivek Prajapat

Exploring technology, AI, psychology, and the inner life of ambitious people

[Vivek Prajapat @vkp_00]

5/5
If you are building an AI product and the backend feels like the weak link.
It probably is.
Java backend dev helping founders fix that layer before it becomes a crisis.
Let's talk. 👇

[Vivek Prajapat @vkp_00]

1/5
You do not need a data scientist to build an AI product.
You need someone who knows what happens when it breaks at 2AM.
That is a backend problem. Not an AI problem. 🧵

[Vivek Prajapat @vkp_00]

4/5
What this layer actually looks like in Java.
Async calls so users never wait on a thread.
Model abstraction so you swap providers without touching product code.
Structured output so AI responses map to real objects not unpredictable strings.
Boring. Critical. Spring AI.

[Vivek Prajapat @vkp_00]

5/5
This is the layer most teams skip.
Retry. Fallback. Queue. Test. Cache. Observe.
Not glamorous. Not a launch announcement.
But it is the difference between an AI feature and an AI product.
Java backend dev thinking about this every day. 👇

[Vivek Prajapat @vkp_00]

1/5
Most AI features work perfectly in the demo.
Then they go live.
The gap between those two moments is almost always a backend problem, not an AI problem. 🧵

[Vivek Prajapat @vkp_00]

4/5
The founders who figure this out early build faster.
Because they stop blaming the model every time something breaks and start asking the right question.
Is the backend built to hold this?

[Vivek Prajapat @vkp_00]

6/6
Building this taught me more than any tutorial ever did.
Real decisions. Real tradeoffs. Real architecture.
This is the project. This is the direction. Collaboration welcome. 👇
https://t.co/OLQpGQSwi2

[Vivek Prajapat @vkp_00]

1/6
I built a full retail management backend from scratch.
Multi-tenant billing, tax compliance, inventory audit trails, CRM with auto-segmentation, supplier management, payment verification.
Here's what actually surprised me building it. 🧵

[Vivek Prajapat @vkp_00]

5/6
Two kinds of people I want to talk to.
Founders building retail tech or tools for non-technical business owners. There's a backend here worth building on.
Developers who want to see what a production-grade Spring Boot system actually looks like. The code is on GitHub.

[Vivek Prajapat @vkp_00]

7/7
Strong backend first. Then AI that makes it usable for anyone.
That's the direction, and the engineer I'm becoming.
Founder building retail tech or AI for non-technical users? I'd like to talk. 👇

[Vivek Prajapat @vkp_00]

1/7
Most retail software falls into two extremes.
Either too complex for a business owner to actually use, or too simple to run a real business on.
I built something in between. Here's what's actually in it. 🧵

[Vivek Prajapat @vkp_00]

6/7
Here's where I'm taking it next.
A business owner shouldn't need a dashboard to understand their business. They should ask "what's about to run out" in plain language and just get an answer.
That's the gap Spring AI closes.