For a long time, the traditional computer science path looked the same: attend lectures, do assignments, build a few projects that only exist on local hosts, upload them to GitHub, and hope a recruiter notices.

In my second year of computer science at NIAMT Ranchi, I realized I was tired of that loop. I had a folder full of half-finished repositories, and none of them had real users. I wanted to build software that actually lived in people’s browsers, solved daily frustrations, and was fully deployed.

So, I set a goal: Ship 10 real, monetized, and useful products in 12 months.

But how does a single student with a busy class schedule build, test, polish, and launch 10 products? The answer is simple: Vibe coding and massive AI leverage.

What is “Vibe Coding”?

Vibe coding is the philosophy of using AI tools (like Claude, Cursor, Gemini, and Codex) as your execution engine, while you act as the architect, product designer, and quality gatekeeper.

Instead of writing every single line of boilerplate express routing or database setup, I focus on:

  1. Identifying real problems that people face.
  2. Designing the product flow and user experience.
  3. Instructing AI engines to generate code within strict architectural patterns.
  4. Testing, debugging, and polishing the result for production.

AI doesn’t replace the developer. It acts as an incredibly fast, highly competent junior developer that works 24/7. It gives one person the leverage of a three-person software team.

From Formula Student to Chrome Web Store

My journey began with NIAMT Racing — the official website for our college’s Formula Student team. This was a classic full-stack project, but it was my first time shipping code that sponsors and event judges would actually look at. That project taught me that a product doesn’t end when the code compiles; it ends when it’s live, fast, and responsive.

After NIAMT Racing, I realized I could solve smaller, highly specific problems by building Chrome Extensions. They are lightweight, run directly in user workflows, and can be published easily to reach thousands of users.

I built:

  • FocusLog: An automatic time tracker that runs locally in the browser with zero server latency, solving the friction of manual timers.
  • JD Saver: A tool that extracts LinkedIn job descriptions in clean PDF/Markdown formats, accompanied by a local application tracker.
  • Media Extractor Pro: A utility that extracts and packages media from social links, including a companion Chrome extension that bulk-downloads Google Maps photos.

Scaling the Vision: 10 Products

As my efficiency grew, so did the list of projects. To showcase the scope of what one student can do with AI leverage, I expanded Garam Softwares’ catalog to include:

  1. Striver DSA Tracker: A customized dashboard to help B.Tech students track their progress through Striver’s A-to-Z DSA course sheet.
  2. PeekPost: A tool that lets content creators visually preview and optimize drafts for LinkedIn and Twitter/X before publishing.
  3. Call.io: A peer-to-peer WebRTC audio and video calling app that requires zero accounts or downloads.
  4. MedScribe AI: An AI-powered medical scribe assistant built for the Gemma 4 Good Hackathon.
  5. VolunteerSync: A collaboration portal matching volunteers with event opportunities.

Every single one of these is live, fully styled, and accessible to anyone.

Lessons Learned Shipping in Public

1. Done is Better Than Perfect

In college, we are taught to worry about perfect academic code. In the real world, shipping beats perfection. Launching an MVP (Minimum Viable Product), seeing how users interact with it, and fixing bugs on the fly teaches you more in a week than writing theoretical schemas for a month.

2. Empathy is the Best Feature

A product is only successful if it solves a real pain point. Striver DSA Tracker works because DSA prep is stressful, and students need visual tracking. JD Saver works because job hunting is messy, and losing job details when postings are taken down is frustrating. Build for real empathy.

3. Vibe Coding Requires Discipline

It’s easy to ask AI to generate code. It is much harder to maintain a clean codebase, write solid CSS patterns, and debug race conditions. As a “vibe coder,” your most valuable skill is reading code and knowing when the AI is hallucinating or building tech debt.

What’s Next?

I’m currently halfway through my 12-month timeline, and the pipeline is full of new tools. By building in public, I’m documenting the successes (and the builds that fell apart) to show other students that they don’t need a corporate budget or a large team to build a software business. All you need is a laptop, an internet connection, and the discipline to ship.

Follow my build log on GitHub or connect with me on LinkedIn to see what we’re shipping next.