My First Year of Engineering: What I Built, Won, and Actually Learned
This isn't a guide to engineering. It's a personal reflection on my first year—what I built, what I won, the mistakes I made, and the lessons that shaped me far beyond the classroom. A year ago, I was looking for an article exactly like this one. Not a motivational post. Not a listicle. An honest account from someone who was actually building things during their first year of engineering in India — the failures, the late nights, the stuff that worked, and the stuff that didn't. I couldn't find o
This isn't a guide to engineering. It's a personal reflection on my first year—what I built, what I won, the mistakes I made, and the lessons that shaped me far beyond the classroom.
I'm Karan Raj — 18 years old, B.Tech Computer Science (AI & ML), NIAT–S-VYASA University, Bengaluru.
A year ago, I was looking for an article exactly like this one.
Not a motivational post. Not a listicle. An honest account from someone who was actually building things during their first year of engineering in India — the failures, the late nights, the stuff that worked, and the stuff that didn't.
I couldn't find one. So here's mine.
Where It Started
In August 2025, I packed my bags, got into my dad's car, and moved to Bengaluru to begin my B.Tech in Computer Science (AI & ML) at NIAT–S-VYASA University.
Like most freshmen, I had more ideas than experience.
I didn't have clients.
I didn't have a network.
I didn't have achievements.
What I did have was curiosity, a GitHub account, and the realization that four years would pass much faster than I imagined.
I decided I didn't want to graduate with only a degree.
Building KĀRYO
Within the first few months of college, Havinash Gangisetty and I founded KĀRYO, a digital agency focused on helping businesses build their online presence. From day one, we've worked as equal founders with a shared goal of helping businesses build a stronger online presence.
One of our earliest projects was developing the public-facing website for ISKCON Visakhapatnam using Next.js, TypeScript, Tailwind CSS, and Framer Motion.
While Havinash focused on the admin dashboard, I worked on backend architecture, client communication, and business development.
Working with real clients in your first semester of college is a different kind of education. Nobody teaches you how to manage client revision requests at 11PM while prepping for a DSA exam the next morning. Deadlines stop being abstract. Feedback stops being optional. Responsibility stops being theoretical. You figure it out, or you don't — and the client notices either way.
BRAVE — Learning to Sell
One of the most valuable experiences of my first year wasn't a product—it was an experiment.
We called it BRAVE.
Instead of waiting for clients to find us, we walked into dental clinics and local businesses, introducing ourselves and pitching website development, appointment booking solutions, and social media management.
Most conversations ended with rejection.
A few became paying clients.
Those conversations taught me more about communication, sales, pricing, confidence, and persistence than any lecture ever could.
Looking back, BRAVE wasn't about earning money.
It was about proving that we could solve problems people were willing to pay for.
Building VoiceRx in 24 Hours
One of the biggest challenges I took on was HackBLR 2026.
More than 2,500 teams participated.
I entered alone.
Within 24 hours, I built VoiceRx, a voice-powered healthcare assistant that used Retrieval-Augmented Generation (RAG) to help users understand symptoms before consulting a doctor.
I built VoiceRx using FastAPI, Vapi, Groq, Llama 3.3 70B, and Qdrant Cloud to create a voice-first healthcare assistant powered by Retrieval-Augmented Generation (RAG).
Nothing worked perfectly.
At around 4 AM, I seriously considered giving up.
Instead, I kept fixing bugs until submission.
VoiceRx eventually finished among the Top 40 teams.
The ranking mattered.
But discovering that I could build an entire AI product alone under extreme pressure mattered much more.
FormPilot — Winning Our First National Hackathon
Later in the year, my teammates Havinash Gangisetty, Saagnik Dey, and I participated in Open Loop 2026, a national-level hackathon hosted in Mangaluru.
We built FormPilot, an AI-powered Chrome Extension that intelligently fills web forms using multiple AI providers.
Built with React, TypeScript, and Manifest V3, the extension focused on making repetitive form filling dramatically faster.
Our approach was simple:
Build a product—not a presentation.
That approach paid off.
We won 1st place.
More importantly, we proved to ourselves that we could compete against strong teams from across India, build a polished product under tight deadlines, and ship something that continued to exist beyond the hackathon.
Unlike many hackathon projects that never move beyond demo day, we continued improving FormPilot after the competition and eventually published it on the Chrome Web Store.
🔗 Try FormPilot: https://chromewebstore.google.com/detail/ffkpekcnpbafklidejfbhinahahaabfi
Beyond Hackathons
Winning competitions was only one part of the year.
I also spent six months serving as a Google Campus Ambassador, helping students discover Google's developer ecosystem while participating in technical events and workshops.
During that journey, I was recognized as the Top Prompt Creator at a Google Pitch Night focused on AI.
I also contributed to GirlScript Summer of Code 2026, where multiple pull requests I worked on were successfully merged into open-source projects.
Outside structured programs, I kept building.
Some projects became hackathon submissions.
Some became client work.
Many simply remained on my GitHub as experiments that helped me become a better engineer.
Not every project needs to become a startup.
Sometimes the biggest return is simply becoming more capable.
The People I Met
The most unexpected part of this year wasn't a product or a competition — it was the people.
I attended TED events at both S-VYASA University and IIM Bangalore, where I sat across from startup founders, researchers, and corporate builders doing things at a scale I was only beginning to understand. I visited the IIT Alumni Centre in Bengaluru for an event and came away with open invitations from IIT-ecosystem startup founders to visit their labs — conversations I never would have had if I'd stayed on campus.
I also met IIT professors and IIM Bangalore faculty whose thinking challenged the way I approached problems.
None of these meetings were planned. Every one of them came from showing up somewhere outside my comfort zone and choosing to start a conversation. That pattern — show up, engage, follow through — turned out to be the most repeatable system I found all year.
What I Actually Learned
- The version that ships beats the version that's perfect.
VoiceRx wasn't polished. At 4AM, with hours left and half the stack broken, I had a choice — quit or submit something imperfect. I submitted. Top 40 of 2,500+ teams. The projects sitting finished on your laptop have never helped anyone.
- Rejection is just data.
With BRAVE, we walked into dental clinics and local businesses cold. No warm intro. No reputation. Most said no immediately. A few said yes and paid us. That ratio taught me more about sales, pricing, and communication than any course could. The nos didn't stop us — they sharpened the pitch.
- Real customers fix your thinking faster than any mentor.
Every client project forced me to think beyond the code. ISKCON Visakhapatnam had real users, real expectations, and real feedback. You stop building features you think are cool and start solving problems people are actually blocked on. That shift in thinking is worth more than any framework you'll learn in class.
- Your network compounds. Your résumé doesn't.
Hackathons led to founders. Founders led to events. Events led to labs. One conversation at a TED event at IIM Bangalore opened doors I didn't know existed. The technical skills got me into the room. The relationships determined what happened after I walked in.
- College gives you one resource nobody talks about — time with low stakes.
You can fail a hackathon submission at 18 with no job, no mortgage, and no employees depending on you. That window closes. I decided the worst possible use of it was waiting until I felt ready.
Looking Ahead
Year one was about proving I could execute — to myself, more than anyone else.
I built products. Worked with real clients. Won a national hackathon. Contributed to open source. Showed up at rooms I wasn't sure I belonged in.
Year two starts with a different question: can I build something that outlasts a semester?
I'm already competing in TakeOver'26, a national hackathon with the finale in Hyderabad this month. The problems I'm working on are bigger. The stakes are higher. The foundation from Year 1 is the only reason I can attempt them.
If you're a first-year student reading this — don't wait until you feel ready. Start building. Ship publicly. Talk to people. The opportunities you're looking for almost always come after you've done the work, not before.
If you're a first-year engineering student reading this, don't wait until you "feel ready."
Start building.
Ship your projects.
Share your work publicly.
Talk to people.
Repeat.
The opportunities you're looking for usually come after you've done the work—not before.
About Me
I'm Karan Raj, one of the founders of KĀRYO and a B.Tech Computer Science (AI & ML) student at NIAT S-VYASA University in Bengaluru, India.
I'm passionate about building AI products, startups, developer tools, and open-source software while documenting everything I learn in public. My goal is to solve real problems through technology, keep learning, and help other students start building earlier than I did.
You can explore my work and connect with me here:
🌐 Website:https://karanrajkr.vercel.app/
💻 GitHub: https://github.com/Karan-Raj-KR
💼 LinkedIn: https://www.linkedin.com/in/karanrajkr/
📸 Instagram: https://instagram.com/karan.rajkr
If you're building something interesting—or you're a student figuring things out—I'd love to connect. Feel free to reach out.
Thanks for reading.