My GSoC Selection Journey | Drona Raj Gyawali

Hi everyone! I am Drona Raj Gyawali, from Lumbini Province, Nepal. I am currently a second-year undergrad majoring in Computer Science at Butwal Multiple Campus. Today, I want to take you through my journey from open-source contributor to being selected for Google Summer of Code (GSoC) 2026 with GreedyBear, a cyber threat intelligence project.

In my first year, I started learning Python with the sole goal of breaking into AI research. Along the way, while implementing ML papers, I realized I was falling in love with backend systems. I found myself happily spending hours deep-diving into system engineering and design. The pull was so strong that I made a bold decision: pause the AI research and go all-in on backend architecture.

I started building, but I quickly hit a roadblock. Back in school or college, if I solved a complex equation, I could ask a teacher to verify it. But here I was writing code with no one to tell me: Is this code actually good? Can this safely run for 15 years? Is this implementation industry-standard? AI is there, sure, but AI sucks at rigorous code review because it’s always context-hungry.

That headache led me to discover the true power of Open Source. The concept pulled me in instantly: you write code for an issue, and senior engineers review it and give you world-class feedback, completely for free.

I started contributing to the Internet Archive/Open Library project and got to work on some amazing issues. Most of it was refactoring backend code to adapt to Ruff linting rules, but I ended up finding and fixing actual bugs.

Through those merged PRs and the guidance of the maintainers, I learned CI/CD, writing robust test cases, shipping production-grade code, and advanced debugging strategies. If you tried to learn this stuff in a commercial bootcamp, they would easily charge you 3–4 lakhs NPR. I was getting it for free from the pros.

This wasn’t an overnight success. Last year, I applied to a cybersecurity organization to work on detecting login anomalies. I didn’t get selected due to a communication gap. However, I still managed to get my name added to the README as a backend contributor because of my work and dedication to the project. Honestly, that felt like a win, and the experience taught me exactly what I needed to fix.

This year, I applied to a different organization with a highly specific domain to truly test myself: GreedyBear.

When I started, I had absolutely zero knowledge of threat intelligence. I began small, adding test cases and refining edge cases. Eventually, I leveled up and merged major features like credential clustering ,ASN endpoints, optimizing code performance and many more thing.

The biggest differentiator this year was communication. I engaged with my mentor from day one. I vividly remember a straight 5-day period where my mentor and I discussed my proposal insanely fast. We were debating architectural details; my mentor was pushing back, and I was in the zone trying to prove my implementation was right. That intense, thoughtful discussion made the proposal rock solid.

I submitted it well before the deadline, waited, and finally got the selection email.

If this journey taught me anything, it’s that communication is key. Any technical problem can be solved with thoughtful, transparent discussion. I’ve already learned an immense amount from my mentor’s PR reviews, sitting at 19+ merged PRs before the deadline.

Tomorrow, the official coding period starts. It’s time to keep writing lean, robust code.

You are still here and then my humble request please follow me on github, I will take that follow count as a token of love :)