How I got Selected into GSOC '2026 from Nepal
This is the overall Journey of mine how my proposals got accepted in the program.
Hi, I am Drona who loves his time to spend on AI and Distributed system also a 2nd year student from Butwal, Nepal. I am currently pursuing my Bachelor of Information and Communication Technology Education (BICTE) at Butwal Multiple Campus, which is affiliated with Tribhuvan University (TU) under the Faculty of Education.
Last year, I thought I knew what Open Source was. I had 26 PRs merged, I was solving bugs, and I was writing complex features. But when the results for GSoC '25 came out, it was a hard rejection. I was crushed, but more importantly, I was confused.
I realized I was treating code like a volume game, when it’s actually a conversation.
This year, I returned to the GSoC arena with a completely different spirit. I joined IntelOwl (and specifically the GreedyBear project), a powerhouse in the threat intelligence space.
I made a conscious choice: Stop chasing the "obvious." In an era where AI can fix a syntax error or a basic bug in seconds, those "easy wins" have lost their signal. Instead, I went looking for the friction. I hunted for the issues that couldn't be solved in a solo, the ones that required deep design discussions, architectural debates, and the courage to defend my logic against the scrutiny of expert mentors.
The result? I ended up with fewer merges than last year org, but the depth was incomparable.
I had Pull Requests with 26+ comments, threads of intense collaboration, feedback, and refinement. Those 26 comments taught me more than 100 solo commits ever could. They proved that the true power of engineering isn't found in your "self-assumptions," but in the collaborative friction that polishes a good idea into a great one.
In a world where LLMs can generate syntax instantly, it’s easy to develop a "collector" mindset, where you feel like you know everything because you can search for everything. But that’s a trap for the ego. I decided that I’d rather take 10 deep, painful learnings from one single PR than a dozen easy wins, and which push me to discuss more and more about problems, in most of the PR my core discussion used to be about latency , db internals and scope, which also helped me to think out of the box.
I stepped into a world of threat intelligence with zero prior knowledge of the field. I didn’t know IP clustering, tpots, honeypots, and many more things about cyber-security and threat intelligence . But because I kept my engineering fundamentals solid, I could learn the concepts on the fly. Today, the boy who knew nothing can lead a discussion for hours on these topics. That is the magic of the Open Source community, it doesn't just build software; it builds people like me.
One more thing: when it comes to your proposal, start long before the official submission window opens. Use that extra time to discuss your ideas openly with your mentors. This collaborative "pre-draft" phase allows you to refine your technical approach based on real feedback, and more importantly, it gives you the genuine confidence that you can, and will, finish what you start. However, This summer, I’ll be building a Event Injector API. It’s a massive undertaking, but I’m going in with a clear lessons I learned.
To sum up, Collaboration is the master key. No matter how hard the problem is, it can be solved through better communication. If you want to be the next contributor, don't just write code. Talk. Listen. Defend your ideas and more important Keep your fundamentals so sharp that they help you when nothing else can.