As I’ve mentioned in previous posts, my internship is built around two main projects. The first is looser and research-based: I’m testing out a theory for an internal AI system the company might use. The second is far more structured and code-heavy — a project broken into tickets, where I complete one element at a time until the final deliverable is done. The problem is that these tickets are packed with coding jargon and buzzwords I’ve never come across before.

My boss’s advice is to leverage AI to teach myself what’s actually going on at Carsdata and what I’m being asked to do. So before I even attempt a ticket, I spend roughly 30 to 40 minutes learning what each component is and what problem the larger project is trying to solve. Then I get to work.

But I can only research and push so far before I hit a wall. A lot of my work happens alongside AI, with the tools acting as a complement to me rather than a replacement — and I don’t have complete faith in them, so I write my prompts with strict guidelines to keep the project from going off-script.

The main AI tools I use for work are Claude and Claude Code, running in tandem. As reliable as they can be, one thing I’ve learned is that Claude — and most AI tools — start to lose the context, the “plot,” the longer you stay in a single chat. I once spent three days on the same ticket, and by the end Claude kept inventing new details about it and new next steps. I’d catch and correct them, but my productivity took a dip every time. So I had it write a context file I could upload into a fresh chat — one that laid out the goal, what had already been done, and what still needed to get finished.

When a roadblock comes up that I genuinely can’t get past, I message my manager. I understand he’s busy and can’t walk me through every task, but my Slack messages often go unanswered for a couple of hours.

I’ve found two ways to stay productive while I wait. The first is to return to the research project — learning new concepts, testing how they might apply, and building the theory further. The second came out of the fact that I have access to such a modern tech stack: GitLab, Snowflake, and Claude Code. I started running my own independent projects to build skills my internship hasn’t touched yet. My role is less about analyzing data and more about building the tools that let others analyze it, so I used that same access to run my own analysis on the company’s resolved ad-to-sale data — and found a relationship between a car’s mileage, how long it sat on the marketplace, and what it eventually sold for.

Finding ways to acclimate to a Czech work environment hasn’t just paid off at the office. The lessons I’ve picked up so far at Carsdata have carried over outside of work, too — onto my first solo trip.

I traveled completely solo to Kraków for the weekend. I’d never really traveled alone before — my last trip was Budapest, where I went with two other kids from my building. But this was my first time as a true solo traveler, and I can confidently say there were plenty of moments where I was completely out of my element: figuring out my transfer trains, finding my hostel, tracking down a decent meal. It all fell on me.

At first, the independence was overwhelming, even a little scary. Having total freedom and agency over an experience like that was almost frightening — no obligations, no one to wait on, just me deciding where to go and what to do. But as nerve-racking as it was going in, I came out the other side with a real sense of accomplishment: in two nights I’d managed to navigate and immerse myself in a city I’d never been to.

I made friends along the way, too. There was a British couple on holiday in Kraków — we traded perspectives on the world, comparing America to Britain, and finding that much common ground so far from home made me feel more worldly than I ever have. And there were the Polish kids I met outside my hostel, where we swapped cultural differences and life experiences. I left Kraków no longer afraid of going out into the world on my own. Honestly, it’s freeing. And learning to manage that independence is what makes the discomfort workable — it’s the difference between being stuck and just being somewhere new.