How an Almaty graduate got into Y Combinator and built a fast-growing AI startup

Self-learning voice AI agents, half a million calls a day, and investment from Y Combinator — the story of a founder from Almaty who built a world-class startup.

Arkady Telegin, San Francisco city, co-founder of Leaping AI, Instagram, LinkedIn

How it all began

I was born and raised in Almaty. Back in my school years, I learned about Y Combinator — they run the news platform Hacker News. I have been reading it since I was 14 and always thought it would be amazing to one day found my own startup and get into Y Combinator. At the time, there was no concrete plan — just an inner feeling that one day I would create something of my own.

After school, I moved to Germany to study computer science for my bachelor’s degree in Munich. Almost immediately, I found a job at a startup where we were building autonomous industrial robots for cleaning airports and other facilities. Later, I worked at several other companies, including Microsoft. Throughout that time, one question kept running through my mind: “Do I really enjoy this?”

At first, I thought maybe I just needed to get into a “cool” company. But once I joined Microsoft, I realized that was not the issue. The team was great, and the product was interesting — we were working on infrastructure for generative models that create images. Despite all of that, I felt like something was missing. That’s when I finally understood: I wanted to build something of my own.

In 2022, I was studying LLMs at university before they became mainstream. Even then, it was clear these technologies could replace call centers, and I saw enormous potential in this space. In the evenings, I wrote code, looked for my first clients, and showed them early versions of the product. When I realized there was strong interest, I dropped out of university, quit my job, and fully committed to the project.

Later, we applied to Y Combinator — the very American accelerator I had dreamed about since childhood — and were accepted into the winter batch. We raised investment and began to scale rapidly.

The team

My co-founder was a university friend who had previously worked at Boston Consulting Group. He was also bored with the corporate environment and decided to join me. He took on the business side — client acquisition and negotiations — while I focused on product development. For almost a year, we worked as a two-person team with no employees.

As the project began to grow, we started hiring. At the moment, we have around eight employees. The team is small by design: fewer people mean lower management overhead and higher efficiency. This is a growing trend in Silicon Valley.

Y Combinator has a section called Request for Startups, where it outlines the areas in which it wants to see new companies emerge. At the moment, they are looking for a startup capable of reaching unicorn status with a team of no more than 100 people.

Target audience

We work with businesses that already have a call center or are planning to build one. These companies provide customer support by phone. Many of them currently have no call centers at all: when customers need help, they have to send an email and wait several days for a response. We help change that.

Our clients include home services companies, retail chains, and distributors. For example, Germany’s largest wine distributor is our client: 60% of its sales happen over the phone, and our AI agents consult customers and take orders.

Achievements

At first, we operated on our own funds and small angel investments. After being accepted into Y Combinator, we received $500 000, and upon completing the program in March 2025, we raised an additional $5 million, with Nexus Ventures as the lead investor.

By the end of the Y Combinator program, our revenue had reached $1 million, and during the 10-week program we doubled it. We ranked among the top three fastest-growing companies in the batch. Paul Graham, the founder of Y Combinator, also invested in us separately, which was particularly meaningful.

Today, around half a million calls pass through our system every day, sometimes reaching up to 50 000 calls per hour.

The challenges

When we started, there were almost no competitors on the market. Today, similar solutions are far more common, and the market is gradually becoming saturated. That is why we focus on brand development and targeted go-to-market strategies — for example, concentrating on a single vertical and capturing as much market share as possible before expanding into adjacent areas.

Future plans

The next step is to add omnichannel support — not only phone calls, but also email, which is important for large enterprise clients. In the long term, we aim to become a leading solution for contact centers: a platform that combines artificial intelligence with human operators.

The main lesson I have learned is to never give up. In the beginning, it is always difficult — failures, doubts, and a lack of resources are inevitable. But if you keep going and do not give up, at some point an opportunity appears, and by then you are already prepared, with experience, a product, and a deep understanding of the market.

As for relocation, building a startup in the United States — especially in San Francisco and the Bay Area — is much easier. The market is enormous, people are open to innovation, and they are more willing to adopt new solutions. In Germany, for example, selling and raising investment was much harder: society is more conservative and skeptical. Building a startup is difficult everywhere, but outside Silicon Valley, it is even harder.