Service
AI agents and LLM apps that work in production
Agents that take actions, chatbots that know your data, AI features inside your product, voice and video pipelines. I build them end to end, with a backend, an interface and a deployment, not as a demo
- Type
- Service
- Published
What can an AI engineer build for your product?
An AI demo takes an evening. A product takes a backend, user accounts, an interface people understand, cost control and a deployment that survives real traffic. I build all of it
Who it's for
- Startups that want AI in the product, not only in the pitch deck.
- Businesses that lose hours every week on routine work: inboxes, leads, reports, content.
- Teams with an AI prototype that now has to become a reliable product.
What I build
- AI agents. Agents that call your APIs and tools and finish a task, not just chat about it.
- Chatbots and assistants on your data. Answers from your documents, knowledge base and product data (RAG).
- AI integration into your product. Generation, search, classification, summaries and recommendations as normal product features.
- Process automation. The routine work that eats your team's time.
- Voice and video. Synthetic voice through the ElevenLabs API, videos rendered in code with Remotion.
- Real-time AI interfaces. Live updates over WebSockets and WebRTC.
How a project goes
- Free call. The task, your data, the constraints and what success looks like.
- Prototype on real data. You see the quality of the answers before paying for the full build.
- Production. API, interface, monitoring, limits and cost control.
- Launch and handover. The code, the accounts and the documentation are yours.
- After launch. Support, iterations or training for your team, agreed per project.
Examples
- Imit, CTO, 2023. A chatbot that reproduced how a specific person wrote, based on their message history. Built on GPT in early 2023.
- AI presenter. An agent that presents a product by voice, switches slides and changes the page as it talks.
- vibecatter videos. Films and explainers built with Claude Code and Remotion, voiced through the ElevenLabs API.
- AI-first engineering. Since 2023 I build everything with AI agents: dozens of hackathons and dozens of repositories, from fintech and web3 to games.
Models and stack
I work with Claude and OpenAI models and use ElevenLabs for voice. The model is chosen per task: quality, speed and price differ for each job. The architecture keeps the provider replaceable, so you are not locked in.
Stack: TypeScript, Node.js, Next.js, React, PostgreSQL, WebSockets, WebRTC, Remotion. Python when the task needs it.
Why hire me as your AI engineer
- Full-stack for about 15 years. AI is one feature in a product; I build the whole product.
- Designer first. About 500 interface layouts, so AI features are easy to use.
- Founder and CTO. I have built products that had to earn, so I care about the result, not the demo.
- Built for load. Game servers with about 3,500 players online and stores that survived launch-day traffic.
Technologies
Questions and answers
How much does an AI project cost?
It depends on scope and speed. After a free call I send an estimate: a fixed price for a defined scope or a monthly rate. Work starts with a prepayment and continues in milestones
How fast can you start?
Right away. Part-time or full-time, depending on the project
Which models do you use?
Mostly Claude and OpenAI, chosen per task. You can switch providers later without rebuilding the product
Is my data safe?
I use the business APIs of the model providers and give the system access only to the data it needs. We agree on the details for your case before the build
Who owns the code?
You do. The code, the accounts and the documentation stay with you
Can you work with my existing team?
Yes: as an engineer who builds the AI part, or as a fractional CTO who leads it
Where are you based?
In Barcelona, Spain (CET). I work remotely with clients worldwide
What language do we work in?
Ukrainian or Russian. In English I work in writing; for calls we can use an interpreter
