Oleg Lykho | Building Aptika

Helping German Mittelstand companies automate regulatory compliance with AI

[Oleg Lykho | Building Aptika @OlehLy]

I’m 17, living in Germany, and I’m building something that I wish had existed when I first started taking sport seriously.

Most of us grew up training in a strange way: random YouTube drills, coaches who don’t have time for everyone, parents who worry about injuries but don’t really know what’s safe or effective. Some talented kids quit because of one bad injury, some never get proper feedback, and a lot of potential just disappears quietly.

I’m trying to solve that.

I want to build a system where any young athlete, regardless of money or location, can learn skills with structure, understand what they’re doing to their body, and see clear progress instead of guessing. A place where training plans are not just copied from pros, but built around your age, your body, your mistakes, and your goals. Where “coaching” is not a privilege but infrastructure.

I’m documenting the whole journey here: from my room, with almost no budget, trying to turn this idea into something real that actually helps people move better, stay safer, and stay in sport longer.

If you care about youth sport, talent that never gets developed, or how we can use technology to build better humans instead of just better feeds, you might find this worth following.

[Dom Kwok | EasyA @dom_kwok]

imagine how many rejections these guys got on their way to the top.

never give up.

[Oleg Lykho | Building Aptika @OlehLy]

The best ideas don’t come in meetings. They come when you’re alone.

With other people around, part of your mind is busy managing them. Alone, with nothing to react to, your brain finally works on the real problem.

Almost everyone who’s done great work knew this

[Oleg Lykho | Building Aptika @OlehLy]

Pick a clear goal. Be willing to be bad at it for an embarrassingly long time. The moment your brain insists there’s no way through is usually the moment you’re closest to one.

Don’t quit there. That’s the most expensive mistake you can make.​​​​​​​​​​​​​​​​

#success #act

[Oleg Lykho | Building Aptika @OlehLy]

People think talent is fixed. It isn’t. The brain shapes itself around whatever you keep doing.

Most of that shaping looks like failure, so most people quit. They decide they “weren’t built for it.” But the brain says that about everything new. It’s not evidence.

[Oleg Lykho | Building Aptika @OlehLy]

A polite smile is the same shape whether someone got it or didn’t. The signal isn’t the praise. It’s whether a stranger can describe the project back to you an hour later.

#startup #pitchadvice

[Oleg Lykho | Building Aptika @OlehLy]

Got home from a pitch night thinking I’d done well, everyone said my idea was great. Then I realized I’m pretty sure none of them understood what it is.

[Oleg Lykho | Building Aptika @OlehLy]

The director thanked me in front of all the teachers today for a site I built. I froze and blushed and mumbled goodbye.

If I’d had a graceful line ready, it would have meant I was imagining the praise while I worked. That I froze is small evidence I was just building.

[hamed @thehamedmp]

1. should you move to us? and why
2. what swedish ecosystem can do?

listen to the full 18 min speech of @paulg at @ycombinator visit to stockholm in apr 2026

[Ronit Pereira @CAronitpereira]

“It takes a lot of hard work to make something simple.”

- Steve Jobs

[Oleg Lykho | Building Aptika @OlehLy]

I went to the AI Tinkerers event in Cologne tonight. I met smart developers, founders, and investors.
I shared my startup idea and people liked it.
I realized I need to practice my English more, but the event was a big success. Can’t wait for the next event!
#aitinkerers

[Oleg Lykho | Building Aptika @OlehLy]

@boardyai Big Tech is great, but at Aptika, we're literally changing how Germany works. We are building autonomous AI agents to solve bureaucracy problem choking the Mittelstand. If you want to engineer the future of invisible compliance instead of optimizing ad clicks, let's connect

[Oleg Lykho | Building Aptika @OlehLy]

@CRudinschi Thanks! That is exactly the hardest part of building this, and it's why generic AI models can't solve compliance out of the box. We handle it by separating the "brain" (the LLM) from the "knowledge" and the "actions”

[Oleg Lykho | Building Aptika @OlehLy]

Germany didn't lose its economic edge to bad engineering.

It lost €146 Billion last year to paperwork.

Here's the exact autonomous AI architecture we're building to replace the bureaucracy: 🧵

[Oleg Lykho | Building Aptika @OlehLy]

Entrepreneurship is a lonely path; time simply flies by between the prime and the dead months.

[Oleg Lykho | Building Aptika @OlehLy]

6/6 Sometimes the legal constraint makes the product better.
Build the oversight into the architecture, not as an afterthought.

What's a technical constraint that improved your product?

[Oleg Lykho | Building Aptika @OlehLy]

We spent 3 months building a fully autonomous tax agent.

Then our lawyer told us it was illegal.

Here's what we rebuilt and why it's actually a better product: 🧵

[Oleg Lykho | Building Aptika @OlehLy]

5/6 The unexpected benefit:
The mandatory HITL gateway is now our primary trust signal with Mittelstand buyers.
"The AI does the work. You retain legal authority" is a stronger sales message than "fully automated."