Обзоры

Arkadiy Telegin: How a Kazakhstani Engineer Built Call-Center Voice AI into Y Combinator W25

Аркадий Телегин: казахстанец, который довёл голосовой ИИ для колл-центров до Y Combinator

Arkadiy Telegin is the co-founder and Chief Technology Officer (CTO) of Leaping AI, a venture-backed startup in Y Combinator’s W25 cohort. The company engineers autonomous voice AI agents that execute inbound and outbound calling campaigns for enterprise contact centers. Within two years of its 2023 founding, the company secured $10.2 million in venture capital, counting Y Combinator co-founder Paul Graham among its direct angel backers.

🎯
RAEM Startup Radar
Leaping AI Startup Profile in Interactive Investor Directory
View on Radar →

From Kazakhstan to Munich: Learning German in 7 Months for Tuition-Free Studies

Telegin was born and raised in Kazakhstan. Upon finishing secondary school, he pursued higher education abroad; however, university tuition across the United States and the United Kingdom was cost-prohibitive. Germany offered tuition-free higher education under one condition: certified fluency in the German language. Telegin mastered German in seven months of intensive self-study and secured admission to the Technical University of Munich (TUM), consistently ranked among Europe’s top engineering institutions. For related insights on Central Asian founders building global AI scale-ups, read our profile on Erzat Dulat and the founding of AI unicorn Higgsfield.

His earliest weeks in Bavaria were spent in a backpacker hostel rather than student dorms, navigating Munich’s notoriously constrained rental housing market. While completing his computer science curriculum, Telegin led the software division at WARR—Germany’s largest student aerospace engineering society—contributing to the autonomous navigation software for the CRATER Mars analog rover. Concurrently, he honed his systems engineering skills during technical stints at Microsoft and autonomous trucking pioneer Plus.

Arkadiy Telegin — Kazakhstani engineer and CTO of Leaping AI
Arkadiy Telegin, CTO and co-founder of Leaping AI.

Corporate Layoffs as an Entrepreneurial Catalyst

Rather than seeking an alternative enterprise engineering position following workforce reductions at Microsoft, Telegin recognized that conversational AI and large language models had matured to solve high-friction commercial problems. Joining forces with co-founders in Munich, he co-founded Leaping AI.

Automating Call Centers: What Leaping AI Actually Solves

Unlike viral generative AI toys generating synthetic imagery or creative copy, Leaping AI addresses an unglamorous, high-volume operational bottleneck: conversational voice agents that replace tier-1 call center operators. Core commercial verticals include residential home services, commercial roofing, insurance claims intake, real estate dispatch, airlines, retail, and telecommunications.

The platform autonomously books customer appointments, resolves routine troubleshooting inquiries, and executes multi-touch outbound call campaigns lasting several weeks. According to internal benchmarks, Leaping AI agents resolve over 90% of routine customer inquiries with positive CSAT scores and autonomously process roughly half of all inbound calls without human agent escalation.

The company positions itself as a self-learning conversational platform that adapts dynamically from live dialogue telemetry rather than relying on brittle, rigid decision trees. While contact center automation is competitive, Leaping AI’s low-latency acoustic models deliver human-like conversational responsiveness.

$500,000 from YC and $10.2M in Total Venture Backing

Y Combinator’s standard investment structure provides $500,000 on standardized terms: $125,000 for 7% common equity plus $375,000 on an uncapped MFN SAFE note. For Leaping AI, YC accelerated broader institutional backing: the team closed a $4.7 million seed round backed by Nexus Venture Partners and Paul Graham, scaling total venture capital raised to $10.2 million alongside Tier-1 investors including Founders Fund, General Catalyst, and Lightspeed Venture Partners.

YC Acceptance: Beating the 1% Funnel

Y Combinator processes upwards of 25,000 global applications per cohort, admitting roughly 250 startups—an acceptance rate under 1%, significantly lower than Ivy League admissions. Beyond upfront capital, YC provides founders with direct Silicon Valley network effects, mentorship from founders who scaled from zero to IPO, and global visibility on Demo Day.

Key Takeaways for Central Asian Software Engineers

Telegin’s trajectory offers an actionable engineering blueprint: achieving tuition-free European education through disciplined language acquisition, gaining hands-on aerospace robotics experience, surviving corporate restructuring, and identifying defensible B2B enterprise problems. Penetrating top-tier venture networks does not require Silicon Valley origins; it demands technical excellence and relentless customer execution. To review other Kazakhstani founders who graduated from Y Combinator, explore our feature on Arman Suleimenov: Y Combinator experience and nFactorial School.

Frequently Asked Questions

Who is Arkadiy Telegin?
A Kazakhstani software engineer, co-founder and CTO of Leaping AI, with prior engineering experience at Microsoft and autonomous driving developer Plus.

What does Leaping AI do?
Builds autonomous voice AI agents for enterprise contact centers, automating appointment scheduling, customer support, and outreach across home improvement, insurance, real estate, and retail.

How much venture capital has Leaping AI raised?
A total of $10.2 million across three rounds, including $500,000 from Y Combinator and a $4.7M seed round featuring Paul Graham and Nexus Venture Partners.

Понравился материал? Поделитесь с другими:

Author picture
Author picture

Похожие материалы

A field researcher installs a camera trap in the Kazakhstan steppe

Support

Rufford Grants for Kazakhstan: How to Apply for Up to £7,000 for Conservation

Rufford Small Grants accepts applications year-round. Check Kazakhstan eligibility, first-grant funding, eligible costs, references ...
Datacenter server rack infrastructure corridor

AI

Meta Details Prometheus Architecture: First Phase of Supercluster Trained Muse Models

Meta’s engineering leadership has published technical architecture details for the initial deployment phase of ...

Silicon semiconductor processor and circuit board hardware

AI

Amazon Unveils TorchNeuron: Native PyTorch Backend for Trainium Accelerators

Amazon has announced TorchNeuron, a native PyTorch backend designed specifically for its custom Trainium ...

Читайте нас в Telegram и Instagram

Подписывайтесь на наши каналы, чтобы первыми получать самые свежие новости казахстанской IT-индустрии, анонсы грантов, стажировок и эксклюзивные интервью.