Linara Bozieva spent 11 years at eBay, working in analytics. In 2024, she was laid off — and at 39, with a family that had recently moved from Switzerland to the US, she faced the same question so many tech workers were asking themselves: what now?
She looked at the job market and saw too many candidates chasing too few openings. "It felt worth it to start building something small, at least for a couple of years, to support my family and me," she said.
A few months later, she launched Ravenopous, a growth marketing agency — but this was no traditional agency. Bozieva built a three-layer AI system with 27 custom agents that handle everything from market research to ad copy generation to revenue conversion. The orchestration layer alone has six agents: a market researcher, data analyst, creative director, finance agent, legal agent, and orchestrator that routes tasks to the execution layer.
"The entire setup costs me under $1,000 a month," she said. "I'm paying for Claude Code, Codex, and ChatGPT. Then I have a few specialized tools like HeyGen and ElevenLabs."
Her analytics background at eBay helped her architect the system. But the AI itself wrote much of the code. "I told the model what I wanted it to do in plain language, the AI produced it, and then I tweaked as needed."
The most surprising discovery? Where the human part is still irreplaceable. AI can process a transcript of a client call, but it cannot fully read the room or identify where the client seems most nervous. "That judgment and client management still have to come from me," she said.
Once a client is fully onboarded, she spends about two hours a week overseeing their account. If she focused only on client oversight, she estimates she could comfortably manage 20 to 25 clients on her own.
"Building something on my own has been incredibly fulfilling and promising," she said. "It changed not just how I work, but what I believe one person can build."