Русская версия

VitaliiAI Product Engineer

I turn ambiguous problems into working AI products

Bring me an ambiguous business or user problem. I work out what is actually needed, define a sensible first version, build it with AI coding agents, test it and ship it — then iterate on real feedback.

Get in touchSee selected work
  • 3 live AI demos
  • 2 production Telegram products
  • 1 public beta
  • End-to-end product ownership

Selected work

Three products, three kinds of proof: AI analysis of business data, AI answers you can verify, and a real business workflow.

RISKOwn product · AI analysis

Company risk analysis from public business data

Problem
“Is it safe to work with this company?” is a vague question. The answer is scattered across registry records, financial statements, court cases and enforcement data.
What I built
Enter a company’s tax ID: the service collects its public registry and business data, a rule-based engine sets a green, yellow or red verdict, and an LLM explains it in a structured report bound to the collected facts.
My role
I built the original product and its core business and AI logic: data integration, LLM requests, the analysis flow and the first UI. The commercial product Legicon later reused this core.
Status
Live public demo for Russian companies. The interface is in Russian; a saved example report opens immediately.
Read the case studyOpen live demo

CORTEXOwn product · Verifiable AI answers

Ask a long video a question, get an answer with timestamped evidence

Problem
Facts in a two-hour interview are hard to find and harder to trust: a summary alone cannot be checked.
What I built
You ask a question; the answer is generated only from retrieved fragments of the recording and cites the moments it relies on. Each citation shows a video frame and the opening words of the fragment, and one click starts the player at that second.
My role
My own project: product and technical architecture, the evidence and timestamp UX, production decisions and releases.
Status
Live demo on a real two-hour founder interview.

The interface and the interview are in Russian; the evidence links work the same in any language.

Open live demo

STOKCENClient project · Business workflow

Customer intake and sales workflow

Problem
Customers ask a sales team to find equipment. Each request has to arrive complete, reach a free manager quickly and turn into a conversation that continues after the first answer, not a lost chat thread.
What I built
In a Telegram bot and Mini App the customer picks the equipment and fills in its parameters step by step. The request is routed to a manager with a push notification; the manager adds a model and price and sends an offer, and the customer gets status updates. The owner manages managers and requests in a Telegram admin panel and sees metrics in a web dashboard.
My role
Built end to end for the client: workflow design, architecture, backend, the bot and Mini App, routing, notifications and the communication flow, testing and deployment.
Status
In production for the client. The Telegram bot is public.

The bot, the Mini App and the diagram labels are in Russian. Telegram is where the workflow runs; the value is the workflow itself.

Open the bot

How I work

  1. ProblemStart from the business or user problem, not from a technology.
  2. UnderstandingFind out what is really needed: users, data, constraints and what “done” means.
  3. ScopeCut it to the smallest version that proves value, and say what is left out.
  4. Working productDesign just enough architecture and build a product people can actually use.
  5. ValidationTest it, ship it to a real environment and change it based on what real use shows.

I build with AI coding agents, mainly Claude Code and Codex. They speed up implementation. I own the product framing, the architecture, the constraints, code review, testing, deployment and the result.

A recent example: agents built five variants of the RISK first screen in a preview lab. I chose one — not the one the agent recommended — and the tested release went live the same evening.

More shipped products

  • MedRecall

    Live demo on synthetic data

    A doctor’s memory of patient calls: a summary across several calls, search for important facts, and every answer checked against the original line of the conversation.

    Open live demo

  • Acoustic

    In production, paid product

    A Telegram bot with a Mini App for recordings: transcription by speaker, an audio player and a summary of facts, tasks and agreements that can be checked against the recording. Users pay for it inside Telegram.

    Open the bot

  • Dating Assistant / AlphaBro

    Public beta

    Helps write a first message from a dating profile and suggests replies from the context of a conversation; the user picks the wording.

    Open the bot

Ways to work together

The same work, in the form that suits you.

  • Team role

    A full-time role in a product team, owning AI features or new products from problem to production.

  • Long-term contract

    An ongoing engagement with a startup or product company, working inside your process and tools.

  • B2B project

    A scoped build for your business: from an unclear request to a working, deployed product.

About

I am a hands-on engineer who takes product ownership. I work across frontend, backend, data, LLMs and deployment, and I am comfortable owning a product alone or in a small team.

Most of my products so far were built for Russian-speaking users and businesses, so their interfaces are in Russian. The work itself — turning an unclear need into a reliable product — does not depend on the market.

Contact

Tell me about the problem, the role or the project. Email or Telegram, whichever is easier.

← Swipe →