# Nozomio's teenage founder turned a failed coding agent into context infrastructure > Source: > Published: 2026-07-25 01:31:56+00:00 [Nozomio](https://www.nozomio.com/?ref=runtimewire) founder Arlan Rakhmetzhanov has spent his teenage years compressing the usual founder timeline: learning to code, leaving school, raising venture capital, moving to San Francisco and discarding an initial product before most founders have held their first engineering job. Rakhmetzhanov, now 19, is the solo founder of [Nozomio](https://www.nozomio.com/?ref=runtimewire), a five-person AI developer-tools operation in [Y Combinator's Summer 2025 batch](https://www.ycombinator.com/companies/nozomio?ref=runtimewire). A [TechCrunch profile published July 24th](https://techcrunch.com/2026/07/24/build-in-public-fail-in-public-what-its-like-to-be-a-founder-under-20-right-now/?ref=runtimewire) described a founder with an all-or-nothing view of the work. He told TechCrunch that he either wins or loses, adding that many young founders share the same mindset. The bravado sits on top of an unusually direct route into Silicon Valley. Rakhmetzhanov started coding at 15 in Kazakhstan, attended summer programs in San Francisco and began contacting founders from [Y Combinator (@ycombinator)](https://x.com/ycombinator?ref=runtimewire) through LinkedIn. One wrote an angel check for his first venture when he was 17, according to TechCrunch. Rakhmetzhanov says he had already launched an education product with 20,000 users and cold-emailed his way into research work at Stanford. He later left high school, traveled to London to raise Nozomio's first institutional capital and arrived in San Francisco for YC as a solo founder. ### A failed agent became the thesis Nozomio began in early 2025 as an AI agent that could read several code repositories, answer questions and help developers understand unfamiliar software. Rakhmetzhanov later gave a blunt assessment of that version in [his seed-round announcement](https://www.linkedin.com/posts/arlan-rakhmetzhanov_announcing-nozomios-62m-round-led-by-activity-7379153211493982208-1mMk?ref=runtimewire): "I failed." The failure produced the idea that now defines Nozomio. While preparing for YC, Rakhmetzhanov found that coding agents routinely missed software versions, selected the wrong frameworks and exhausted their context windows. He shifted the product away from acting as a standalone engineer and toward supplying other agents with current, source-backed information. The resulting product, [Nia](https://www.trynia.ai/?ref=runtimewire), indexes code, documentation, PDFs, datasets and internal files. Developers can connect it to agents through an API, command-line interface or Model Context Protocol server. [Nia's documentation](https://docs.trynia.ai/welcome?ref=runtimewire) describes a retrieval layer that continuously monitors information and hands relevant material to coding agents such as Cursor or Claude. That positioning lets Nozomio sell infrastructure without asking developers to abandon their existing editor or agent. It also places Nozomio underneath products controlled by larger vendors. Nia has to remain useful as model providers expand their own search, memory and repository tools. Rakhmetzhanov's thesis has widened since the original coding product. Nozomio now calls itself a product lab focused on context across multiple domains. Its [current product lineup](https://www.nozomio.com/about?ref=runtimewire) includes Nia and [Folk](https://www.folk.com/?ref=runtimewire), a personal agent designed to operate through messaging services. Nozomio is applying the same underlying premise to software and consumer products: an agent becomes more useful when it can retain relevant history and retrieve accurate information at the right moment. ### Investors backed speed before public metrics In May 2025, when Rakhmetzhanov was 18, Nozomio announced an $850,000 pre-seed led by [LocalGlobe](https://localglobecapital.com/?ref=runtimewire). The round also included No Label Ventures, Andrena Ventures, Ventures Together, Eurasian Hub Ventures and Artificial Societies, according to [Startups Magazine](https://startupsmagazine.co.uk/article-18-year-old-founder-raises-850k-build-autonomous-ai-engineers-dev-teams?ref=runtimewire). By October 2025, Nozomio had announced a [$6.2 million seed](https://www.linkedin.com/posts/arlan-rakhmetzhanov_announcing-nozomios-62m-round-led-by-activity-7379153211493982208-1mMk?ref=runtimewire) led by [CRV](https://www.crv.com/?ref=runtimewire), [BoxGroup](https://www.boxgroup.com/about?ref=runtimewire) and LocalGlobe. YC and [20VC Product](https://20vc.fund/portfolio?ref=runtimewire) participated, alongside angels including Paul Graham, Hugging Face co-founder Thomas Wolf and Spotify co-president Gustav Soderstrom. Nozomio has not disclosed a valuation. The investor list reflects a wager on Rakhmetzhanov's pace and the persistence of the context problem. Public evidence for Nozomio's commercial traction remains thin. Nozomio has not published verified revenue, annual recurring revenue, retention or paid-customer figures. Nia does have a visible path to revenue. Its pricing begins with a free tier, followed by a $15 monthly developer plan, a $50-per-seat team plan and a $99-per-seat business plan. Enterprise pricing is negotiated. Nozomio has also published internal benchmarks claiming that Nia improved Cursor's performance by 27% after indexing external documentation. A later [Nozomio benchmark](https://www.trynia.ai/blog/nia-context-augmentation-benchmark?ref=runtimewire) reported lower error rates on tasks involving new and beta APIs. Nozomio built and judged those evaluations, so they show how Rakhmetzhanov wants Nia measured rather than providing an independent assessment of its performance. The public code footprint is still modest. [Nozomio's GitHub organization](https://github.com/nozomio-labs?ref=runtimewire) contains SDKs, documentation, a command-line tool and the original open-source Nia repository. The repositories make integration work visible. ### Building in public accelerates the clock Rakhmetzhanov's story captures what AI has changed for young founders. Models and coding agents make it possible for one technical founder to prototype, operate and discard products quickly. Distribution, reliability and revenue still decide whether that speed creates a lasting business. The same tools that help young founders ship also raise investor expectations. A teenager who can build a working product in weeks can be expected to find customers, publish metrics and pivot at the same pace. Social platforms expose each attempt, including the abandoned versions that earlier generations of founders could have tested privately. Rakhmetzhanov has embraced that exposure. He has described the first Nozomio product as a failure, shared the reasoning behind the pivot and continued broadening Nozomio's product scope. That openness creates a recognizable founder narrative. It also gives customers and investors a public record against which to measure each new claim. Nozomio now needs to turn Rakhmetzhanov's speed into a product advantage that survives changes in models and developer interfaces. Nia must provide better retrieval, fresher source material or easier integration than the context features bundled into the agents it serves. Folk extends the bet into personal data, where privacy, consent and trust carry greater weight. Rakhmetzhanov's age draws attention. Nozomio's next phase will be judged on familiar evidence: repeated use, paying customers and whether its context layer remains valuable as the AI platforms around it improve.