"So, what did you do in the last two years?"
People keep asking. On a CV, it can look like not much happened: no new employer, no new city, no title upgrade.
While friends were collecting promotions and cleaner career lines, I was back in my childhood room, working out what kind of life I actually wanted to build.
Here is the short version: I spent a year at home finishing my master's, then a year re-entering the tech world and teaching myself to build with AI. Along the way, I said no to junior VC, a PhD, and several startup roles that would have looked great on paper. I am now building toward founding through my own projects.
That is the resume version. The real version is messier and more interesting.
At some point, I stopped optimizing for legibility and started building the skills I actually wanted.
The year at home
After a semester in Istanbul, I came back to Germany expecting a short transition: start the internship I had lined up, write my thesis, finish my master's, and continue from there.
Then life got very private very quickly. My family needed me in a way that did not fit neatly into a career plan. The internship stopped being the right priority.
So I stayed home.
I had already stretched my master's longer than planned. Pausing again felt strange. I knew why I was staying, but I still caught myself checking whether I was falling behind.
The tidy career sequence (thesis, job, promotion, next city) stopped moving in a straight line. Instead, the year took on a more basic rhythm: show up, take care, keep going. I cooked. I sewed. I painted and built bits of furniture. I ran and took long walks in the forest. I gardened with my parents. I drove to hospitals, filled out paperwork, and listened.
I wrote my master's thesis from the same desk where I had done my homework as a kid. I also read about AI and startups obsessively, because that thread never broke. I filled notebooks with ideas I had been carrying for years.
I stayed off Instagram for seven months. Not as a statement. I just had no desire to watch other people's curated lives when the real one in front of me finally had my full attention.
There is a version of my twenties where I would have white-knuckled through that period and punished myself for not progressing. I am glad I did not live that version.
Somewhere in the quiet, a question came back that I had not asked myself in years: who am I without the labels? Grades, logos, acceptances, titles had always made the answer easier. With those stripped away, I had to look elsewhere. What remained was steadier than I expected. I still reached for long-form content over short. I still cared about ideas without deadlines or applause. I still liked building and fixing things, even in the most offline sense.
On paper, that year reads like a pause. In my actual life, it was one of the most concentrated years I have had. I learned to enjoy my own company, rediscovered old hobbies, picked up new ones, and was present for the people who mattered most. No one gives you a stamp for that, but it quietly changes how you walk into rooms later.
Re-entry
By the time things stabilized and my thesis was done, I had been out of the visible DACH startup world for about a year and a half. My peers had continuous work experience. I had a master's and a quiet year offline.
Re-entry was deliberate. I reached out to people. I went to Munich at least once a month, stayed with friends, did coffees, coworking, and events. I took ecosystem trips to Berlin and other cities. I went to hackathons and AI meetups. With friends, I ran mini hackathons just for fun. One Tuesday, we built a social media app exclusively for dog content, on the grounds that dog videos are good for your mental health. I let my mind run in every direction.
At first, I felt oddly in between: older than some very young builders who were already shipping, still early in career terms compared to my peers. Inside the builder world, that mattered less. Everyone was experimenting, learning, shipping, changing direction anyway. Outside it, the path was harder to explain.
People understand a student. They understand an employee. They understand a founder once there is a company name attached. The in-between state is socially expensive. You keep having to translate yourself before anyone can decide what they think.
Eventually, I stopped asking whether the path made sense to every room I entered. The better question was whether I was becoming more capable at the work I actually wanted to do.
The paths I did not take
Several tracks were open during this time. I had conversations and some offers. I said no to all of them.
Junior VC was the intellectually seductive one. I still think venture is one of the most interesting games we have invented: deciding which futures to fund, allocating capital against uncertainty, thinking in systems instead of single products. But when I looked honestly at the day-to-day, I saw a role built around sourcing, evaluating, memo-writing, and understanding companies from the outside. I could imagine returning to venture later, but I wanted more operating weight of my own first.
A PhD would have fit neatly into my family's worldview. Education was never just a credential in our house; it was treated as a duty and a privilege. Part of me still loves the idea of earning a doctoral degree. But I was not trying to become an academic. I respected the path, but I could feel the difference between pursuing knowledge and using another credential to postpone a harder choice.
Startup roles were the hardest to decline because some of them were genuinely exciting: good teams, interesting projects, cities I could imagine living in, environments I would have learned from. That made the no harder, not easier. But most of the roles still put me back into one box: operations, sales, growth. I am good at those things, and I respect them. I just did not want "non-technical" to become my permanent label at the exact moment AI was making that distinction less interesting. A few opportunities dissolved on their own too: teams pivoted, priorities shifted, the ground kept moving.
Looking back, the common thread was not that these paths were bad. They were good paths. That was the hard part. Each one made me easier to explain. None of them made the work feel more like mine.
Learning to build
I have wanted to found something for a long time. For years, I told myself there was a sequence: first collect enough experience, then start. There is truth in that. Many successful founders are older, backed by years of industry work. But AI changed the equation for me. Becoming technical suddenly felt both more urgent and more possible. I did not want to stay on the usual path while the tools were changing underneath me. I wanted to be close to the work, early enough to learn by doing. At some point, "later" stopped sounding responsible and started sounding like avoidance.
I was not starting from zero. My degree had real CS in it: algorithms, databases, computer architecture, a search engine I built from scratch in Java. But I had not lived in a repo for a while. Operator and investor work had shaped a different instinct: who cares, why now, what changes behavior, what makes this worth building at all. The impostor feeling was real, and closing it felt overdue.
I built prototypes, worked through tutorials, and wrapped models around workflows I actually used. I could glue things together fast. But speed without depth felt unsafe. Software you cannot fully control is software where you do not know what you have exposed. I did not want to be fast in a way that created landmines for users, teams, or future me.
It was humbling. Python was less forgiving than pitch decks. Retrieval pipelines did not care about my intentions. Error messages do not bend for ego. That is why I signed up for an AI engineering course. The goal was simple: understand enough to be part of the conversation, not just adjacent to it. The tools move fast, but the course gave me a foundation: retrieval, evals, agent workflows, and a clearer sense of where demos start breaking when they become real products.
My childhood room was not a fallback anymore; it became a private lab. I had my setup, slept well, ate well, trained almost every day, and could work deeply without constantly performing momentum. I was still taking trips, doing calls, building with friends over Zoom, and showing up in other cities, so it did not feel like disappearing. It felt like choosing quiet as the default and movement as the exception.
That was a privilege. Not everyone has a family home to return to, parents who make space, or the financial runway to keep burn low while figuring things out. I had that, and I tried to use it seriously.
The tax was real too: fewer spontaneous coffees, fewer founder dinners, friends further away, less serendipity, and doubts I had to sit with alone. The hardest part was not knowing if any of it was going anywhere, while still choosing the same direction without much external confirmation. Some mornings it was just me, my laptop, my standards, and the question of whether I still meant it when nobody was clapping. For that phase, the quiet gave me room to learn without constantly explaining myself.
But preparation has its own gravity. You can keep learning forever and still avoid the harder work of choosing a direction. At some point, the question I had been circling finally sharpened.
Am I learning to build, or learning instead of building?
That question changed the work. I started moving from general preparation into sharper exploration: which problems kept pulling me back, which spaces I might actually want to found in, and what kind of people I might want to build with. I had conversations. I tested ideas. I paid more attention to where there was real pull rather than abstract interest.
Building alone was still a choice, not just a side effect of being home. It is easier to bring the right people in when there is already something to react to: a direction, a rough product, a reason to say yes. I learned I can go deep on my own. I also learned I do not want to stay there forever.
What I shipped
Shipping changed the feedback loop. You can have a beautiful theory about a market, a product, a user, or yourself. Then you ship something and reality replies. Sometimes kindly. Often not.
I kept noticing the same patterns: context scattered across too many places, trust missing from AI workflows, systems too tangled for the people who needed them most. I became more thesis-driven. I was trying to understand which problems matched my values, my curiosity, and the kind of life I could imagine sustaining. The projects I stayed with longest were the ones where the problem felt personally real.
I built a German Parliament Monitor because so much gets debated in public, yet there is no good way to ask your own questions about what is actually happening. The fundraising copilot came from a different kind of recurring question: friends and founders asking which VCs might be a good fit. I also ran research sprints around dating, govtech, home cloud, utility apps, and community products, while staying close to conversations with friends and founders about strategy, product, fundraising, pivots, and what to stop doing. Some of those became prototypes. Others stayed explorations, or only needed to be explored far enough to teach me why they mattered or why they did not.
I built Mini-Me because I kept running into the same problem in my own work: context was everywhere, but never where the agent needed it. My voice lived in one place, project decisions in another, and the reasoning behind choices disappeared across chats. So I started building a private context layer that carries my memory, writing rules, source material, decision frameworks, and project handoffs across tools. Less "second brain," more working context that travels with me.
I built Just Move To Europe because knowledge about European tech ecosystems is scattered everywhere: chats, Notion docs, half-dead blog posts, LinkedIn posts, X threads, and people's heads. The first version started from a simple observation: people do not just need relocation advice. They need trusted local context: which communities are alive, where to work, what city actually fits them, and which practical links save hours of searching. It reached 8,000+ visitors and 26,000 page views in the first four days, and has since expanded to 12 city guides. More importantly, it made a pattern visible: context is only useful when it is current, local, and actionable.
I also wrote "The Single-Use Software Era," an essay about where I think software is going as AI makes more of it easier to generate. It helped me separate tool noise from the deeper structural bets I actually want to spend time on: context, trust, memory, reviewability, taste, and interfaces that give people time back instead of pulling them further into screens.
The surfaces were different, but the pull was the same: making powerful technology easier to understand, trust, and live with. Not AI as magic dust on top of everything. AI that earns its place because it makes a system clearer, safer, faster, or more human.
What I know now
The anxiety of falling behind is mostly gone. Not because I caught up. I do not think there is such a thing. But because I stopped measuring myself against timelines that were never mine.
The last two years gave me recency and proof. I want to work where AI, product, and people meet: practical enough to ask who cares, technical enough to build, and clear enough to explain, sell, or teach.
What excites me now is the kind of work I want more of: harder problems, higher standards, sharper people, and the chance to build something excellent enough to matter.
Staying sometimes felt like failure. Like I should have been somewhere else, doing something more impressive. But staying home became one of the best decisions I made in my twenties. I showed up for my family when it mattered. I rebuilt my own foundation without rushing. And I enjoyed it. For me, time with family was not wasted time.
Confidence did not come first. It came from small pieces of proof: shipping something, surviving the doubt, doing it again, and becoming a little harder to shake each time.
Some days, I still feel the old reflex: wanting a clean title before I feel allowed to explain myself. But the people I respect most do not look at CVs alone to understand someone. They look at what you have built, what you have shipped, the risks you have taken, and who you became along the way.
So, what did I do in the last two years?
I stayed when staying mattered. I learned what I needed to learn. I said no to good-looking paths that were not mine. I built enough proof to stop needing a cleaner title to explain the work.
I chose this. I am glad I did.
I am building around AI, context, trust, memory, and products that make powerful systems easier to live with.
If you are building, investing, or thinking seriously in that space, I would love to talk.
I keep a running list of what I am building here.