Post Summer Internship Check in (My bad for disappearing)

Hey everyone, it’s been a while, got caught up with work and life you know how it is. Interning, interviewing, and trying to enjoy my last “free” summer. I’ll give y’all a quick recap before we dive into the details.

Started up my 3rd internship as a data engineer; got to work on some LLM, AI stuff, I guess, since that’s all the rage now. Realized that the time frame for me to move out of my parents’ house is closing in; excited, also a bit nervous. Also realized that I really fucking hate corporate life, and can’t see myself as an engineer.

My summer doing a bunch of AI slop

So this summer I went back to a company I interned at previously as a data engineer intern; honestly, I had no idea what kind of work they would have for me tbh: traditional data engineering with like Spark, Kafka, managing DAGs in Airflow. Maybe being a machine learning engineer implementing traditional ML methods for forecasting.

Well, a mix of those actually; I got to actively work with RAGs (retrieval-augmented generation), LangChain/Smith, and Kafka. I think it was a good final progression throughout my journey as an intern, now finally taking a project all the way from discovery and research to a poc all the way to deploying it for production. Even though I’m no longer interning there, the hope was that if my project was successful for the first team it was meant for, then it could be expanded to others as well, increasing its utility. Which I think is pretty nice: an intern project actively improving quality of life for various teams from different departments.

To give a brief overview of what my main project was: basically, we have an existing RAG-based policy querying system, and in its current form, for the metrics we wanted to track, the LangSmith dashboard wasn’t sufficient, and the manual review process took 3-4 days at a time, involving multiple people obviously a major time sink. So I was tasked with developing an automated evaluation pipeline for this to make the process faster and scalable, hopefully good enough for use on other teams and departments’ agent evaluation systems. The tech stack was pretty straightforward: AWS S3, Glue, LangSmith, the agentic api for the policy querying system, and of course Python with some standard libraries.

Well, the project was a success. I was able to get manual review time down from 3-4 days to <10 minutes, so big win for me and brownie points to show off to my manager. Finishing the project and looking at the experience now, it really feels like a full-circle moment, going from my first internship being confused and intimidated by the scale of a company’s codebase to actually being able to add meaningfully and see my work actually be used by others.

Sadly, no return offer again- bummer, I know, but thankfully we hedged against this, so no skin off my back. Though it would have been nice

This AI stuff is actually getting pretty good ngl

Over the course of the summer, my company was really pushing us to use tools like Copilot and opencode with Claude Sonnet 4.6, Opus, GPT 5.6 just all the popular models for our day-to-day work. I will say I took advantage cause its not my money being burned, so why not.

I am honestly blown away at how much better these models have gotten. I haven’t really dabbled with anything beyond the free tier with Gemini and GPT. I might be a bit of a doomer saying this, but these models can already do all the work that the vast majority of us junior engineers are being tasked with; they just need someone who understands what the intent, architecture, and bigger picture will be. Like, it was very obvious that if my mentor had just used Claude to build my project, he would have been done in a max of a week, granted he had a better idea of the scope for the project before I started on it.

Yes, yes, I know the discourse now is that understanding architecture is more important than ever, but it feels like a mix of both career and identity crisis I see for a lot of my new grad peers in the field. Personally, I never found programming enjoyable more of a tool to work on the problems I do find interesting, namely statistics. I do understand the people who came into this field loving the tool of programming and feeling lost, not being able to do it. But I think they will find greater enjoyment being able to think about the larger picture and actual system design rather than programming, since this seems like the direction we are heading, but who knows, really. But these AI tools are here to stay so we should use them but still be wary, since none of us can stop change.

Realizing that I may not be cut out for a career as an engineer

After working these last few internships and finishing college, I can’t see myself working as an engineer or really in corporate long term without at least trying something different. I think part of it has also been somewhat subconscious; seeing my family run their own business my whole life, it made the whole corporate life seem like something that wasn’t going to cut it.

I have an idea for a start-up, but I think I’m gonna take some more time to save money and gain experience as an engineer before going all in. But trust it’s either going to succeed or fail horrifically, as all businesses do. I will talk about maybe more in a future blog once I have started working on it, but ya, I think I’m gonna give myself maybe max 2 years in corporate before going towards a start-up and seeing what happens.

Planning my next adventure and trying a new format

Now that I am back to being an unemployed chud waiting to start my full time I plan to finally do some traveling. I got a few places on the list: San Diego, Austin, and Philadelphia. The main reason is to visit friends who have moved to these cities either for post-grad or are currently in school there. Might as well make a few more carefree memories before locking in. I have also talked with a friend about making videos of me just talking about topics I find interesting or just yapping about what’s going on in my life, and maybe record some gameplay. I was thinking maybe a podcast type of thing, but there are too many of those out now anyway, so maybe just videos with gameplay in the background for now. Hopefully in the next month or so I try making my first video and see how it goes.

That’s all for now. I will try be trying to post again in a week or two rather than after the summer. But ya thanks for reading, and talk to everyone on my next post.