Looking Back on My First Internship

My short reflection after my first internship.

It started in November 2025, right after the end of Semester 2 and my exams. I took a quick trip to Korea for about two weeks, then came back for a summer internship.

Through CSIRO, I had the chance to work as a research intern at Pawsey Supercomputing Centre.

In 2025, I tried my best with job applications. I really wanted to gain experience related to what I had learned so far. It’s just rewarding, and the thing I realised is: if I can help someone with my knowledge or skills, like when I was doing my capstone and helping out FOWI (I actually need to write another diary about my capstone, what went well, what went wrong, etc.), it’s just so fulfilling. I could genuinely feel happy in those moments.

I actually counted, I sent about ~50 different applications. I got lots of “no, sorry” emails, and I just kept going. Working, studying, attending events and webinars, updating my resume, getting an AWS certificate, etc. That little bit of effort really did work, and I got the call from Pawsey. Looking back now, I feel like I’m really lucky to have gotten this opportunity.

It was a really well-organised system: one week of intern training, getting paid while learning, and learning so many new areas too. HPC, computational chemistry, agentic AI, and more. It’s such a valuable experience.

They also ran workshops for interns, including sessions on how to communicate scientific research effectively. I’m really grateful for opportunities like this where they invest in students. It makes me feel motivated to work even harder.

I met a lot of people too, and it’s a memory I won’t forget.

And with the CSIRO benefits, I took courses on O’Reilly like agentic AI and software architecture, and they helped me a lot.

I’m from a computer science background, not software engineering, so I lack a bit of software testing, architecture, and that kind of stuff. I could use this chance to catch up on the theory.

Now that Claude Opus 4.6 is out, and you can even control apps on your computer with OpenClaw, my conclusion stays the same: it is still not enough for fully effortless vibe coding. That is what I learned while building this chemistry agent.

I do not think software engineering is dead. I understand that industries will not hire many entry‑level roles anymore, since far fewer people are needed to write code. It is like a farm that once needed one hundred workers but now only needs ten because the tools have improved. With rapid AI progress, especially coding agents, writing code is starting to feel as easy as using Grammarly to write an email. But core areas such as software architecture and system design remain very valuable.

The singularity, the moment when AGI or something similar surpasses human intelligence, would be the point where human software engineers are no longer needed.

But will that moment ever arrive? Someone still needs to take responsibility, and it is unclear how a non‑human entity would handle legal accountability. Policies tend to move slowly as well.

These ideas sum up my recent thoughts. Now it is time to focus on my 2026 honours.