Q3 · 2026 · applications open · 3 of 8 slots open · free process review →
ZININ × MINAKOV
Online · 1:1 and team · RU · CIS · EU

Learn to work with AI — built around your company

A personal program for mastering AI tools — built around the way your company actually works. We take you from "none of this makes sense" to "I build the agents and tools myself, and they do the work alongside my team."

Get a free process review → See formats and pricing
✓ on your company's real work ✓ code and docs stay with you ✓ no black boxes
Q3 Practice status live
applicationsopen
open slots3 of 8
formatZoom · your processes
first call30 min · free
Personal200,000 ₽/mo Teamfrom 400,000 ₽/mo Hackathon ★~1,000,000 ₽ Turnkey agentfrom 90,000 ₽ Ongoingfrom 500,000 ₽/mo
hands-on delivery  ▸  skill handover  ▸  your team runs it  ▸  no black boxes  ▸  code and docs stay with you  ▸   hands-on delivery  ▸  skill handover  ▸  your team runs it  ▸  no black boxes  ▸  code and docs stay with you  ▸  
01Three standard routes

Courses, agencies and hires share one flaw: the expertise leaves with the money

ROUTE 01THE SKILL FADES

Course

The team learns away from real work. People go back to their desks, and two weeks later the skill is gone: nowhere to apply it and no one to help along the way.

ROUTE 02EXPERTISE WALKS OUT

Agency

The contractor takes the project, burns a serious budget, ships the product and leaves. The expertise leaves too. You keep a result nobody can develop.

ROUTE 03SINGLE POINT OF FAILURE

Hiring an AI engineer

A rare specialist on a market salary. The company builds its practice around one person who gets more expensive every quarter — and who, on average, leaves before the expertise reaches the team.

All three end the same way: the knowledge sits with an outsider, and the company is back to depending on the job market or the next budget.

02What we do

The practice is delivery plus skill handover. Both parts are mandatory

+ How the practice works
  • +We work hands-on inside your team. We build the tools together with your people, so the team sees how it works from day one.
  • +We hand the skill over as we go. Training runs on the company's own tasks; by the end the team works with the tools on its own.
  • +We leave documentation and working code. Configured tools, readable code, documentation in plain language. No black boxes.
  • +We stay in touch after delivery. Alexander supports the team, helps adapt the tools and makes sure the practice keeps growing.
✓ Principles
  • ✓Training happens only on the company's tasks, on real processes from day one.
  • ✓We leave only once the team works with the tools on its own. That is the completion criterion.
  • ✓The goal of the practice is to make ourselves unnecessary: the expertise stays inside the team.
  • ✓We promise no numbers — results depend on your processes and your team's readiness, and we say that at the first call.
03What we work with

The tools we teach

All of it is proven in production on real projects. We pass on the version of the knowledge that holds up in live processes.

01 / 06

Claude Code

$ claude code --help
anthropicsubagentsagent-teamshooksMCP
02 / 06

Codex

$ codex run --task ...
openaigptcli
03 / 06

Hermes

$ hermes serve --provider local
open-sourcemulti-llmself-hostedMCP
04 / 06

Cline (Open Claude)

// extension · cline
vscodeide-agentn8ndev
05 / 06

opencode

$ opencode --model ...
open-sourceself-hostedMCPsst
06 / 06

Production stack: MCP, n8n, Langfuse

$ docker compose up mcp+n8n+langfuse
MCPn8nLangfuseUmamiproduction
04Where it runs

Where the AI runs: global, ru

We work with international models and with Russian corporate environments alike — matched to the task, the data and the requirements.

// GLOBAL · international models
AnthropicClaude · Claude Code
OpenAIGPT · o-series · Whisper
GoogleGemini · Veo
MiniMaxvideo · voice
Mistralopen-weights
MetaLlama · open-weights
// RU · Russian models and corporate environments
YandexYandexGPT · YaSpeech
GigaChatSber · GigaCode
T-Bank AIT-One · T-Pro
MWS GPTMTS · corporate environment
05Practice stages

Delivery runs in stages, and at every stage the team works hands-on

01DIAGNOSTICS

We map how the team actually works. We look at the real processes, find the routine work and draw an automation map.

02DELIVERY

We build the tools with the team. Agents, pipelines, bots, integrations with CRM and channels; your people work alongside us and see how it is built.

03HANDOVER

We hand control to the team. The team runs it, we stay nearby as support: answering questions, helping with the first changes.

04INDEPENDENCE

The team grows the practice itself. It owns the tools; Alexander stays in touch so the practice keeps growing.

06Who it's for

For those who want a real result inside the company: working tools in the hands of a team that can develop them

01

A founder or executive with concrete problems — processes that eat time or rest on one person.

02

A department head who needs to rebuild how the team works: marketing, sales, support, analytics.

03

A company that already tried courses, subscriptions or an agency, but the in-house expertise stayed shallow.

04

A team with no technical specialists — we set up tools for real tasks and teach people to work at the level they need.

07What the company keeps

After every stage the company holds concrete artifacts — and the team knows how to work with them

01

Working AI tools — agents, pipelines, bots, integrations inside the company's real environment.

02

Code the team understands — documented, and safe to change on your own.

03

Process documentation — how each tool works and how to adapt it.

04

A trained team — the main artifact: people who can maintain and develop the tools.

05

A map of next steps — what to automate next and in what order.

08The core choice

The difference between a course, an agency and a practice is where the expertise ends up

Course Agency Zinin × Minakov practice
Where the team learnsIn a training environmentNowhereOn the company's real tasks
Who does the workThe team, after the trainingA contractorUs and the team together
Where the expertise goesStays shallowLeaves with the contractorStays inside the team
What the company keepsCertificates and loginsA product with no supportTools and the skill to develop them
Dependence on usRemainsHighMinimal — that's the whole point

Course

Learning in a training environment · the expertise stays shallow · the company keeps certificates and logins · the dependence remains.

Agency

The contractor does the work · the expertise leaves with them · you keep a product with no support · the dependence is high.

Zinin × Minakov practice

Learning on real tasks · we build it with the team · the expertise stays inside · the company keeps the tools and the skill to develop them · the dependence is minimal — that's the whole point.

09Formats

Five formats. Fixed prices.

We'll pick the format for your task and team size at the first call. All materials, code and documentation stay with you for good.

Personal
200,000 ₽ / mo

One-on-one work with a founder or executive: their tasks, their tools, their pace.

  • ✓One-on-one with the engineer
  • ✓Tools for your own tasks
  • ✓Support between sessions
Let's discuss your format →
Recommended
Team
from 400,000 ₽ / mo

Delivery into a team or department's processes, with skill handover. Base: up to 4 people and up to 4 sessions a month.

Size it for your team
Sessions per month
People in the group
4×4 = 400,000 · 8×4 = 640,000 · 4×8 = 520,000 · 8×8 = 760,000 ₽/mo
Let's discuss your format →
Flagship format
Corporate hackathon
~1,000,000 ₽

Two days in which the team builds working AI tools for the company's own tasks with its own hands.

  • ✓2 days of intensive work
  • ✓The team builds the tools itself
  • ✓Working prototypes at the end
Let's discuss your format →
Turnkey agent
from 90,000 ₽

A ready agent for one specific process: sales, support, external communications, marketing. The final quote depends on your processes.

  • ✓One process — one agent
  • ✓Built into your channels and CRM
  • ✓Handed over with documentation
Let's discuss your format →
Ongoing
from 500,000 ₽ / mo

Ongoing support and growth of your AI setup: we adapt and add tools, and keep the team's toolkit current so they work on an up-to-date stack.

  • ✓Regular sessions with the team
  • ✓New tools as you grow
  • ✓Alexander works with the team continuously
Let's discuss your format →

Prices are fixed and identical across the Zinin & Partners network.

10What we automate

Four areas where AI starts paying off fastest

Sales team

Sorting inbound leads, drafting emails and proposals, filling in CRM records, deal reminders.

Customer support

Answers to routine questions, ticket routing, draft replies for agents, a knowledge base.

External communications and marketing

Copy and newsletters, publishing prep, feedback analysis, channel reports.

Agents for one job

A dedicated agent for a specific process: documents, analytics, internal policies — built for your environment.

11Common questions

Answers to what people ask most

How is the practice different from a course?+

A course teaches on textbook examples; the practice teaches on real tasks. The team comes out with a working stack and an understanding of how to maintain it.

What if we have no technical team?+

That is a common situation. We build tools to match the level of the people who will use them; writing code is not required.

How long does delivery take?+

It depends on the format and your processes. At the first call we go through 2–3 of your processes and give you real timelines.

Will it all keep working without you?+

That is the main completion criterion: we leave when the team maintains and adapts the tools on its own. The code and documentation stay in your environment.

What stack do you work with?+

Claude Code, AI agents, Python, n8n, Telegram bots, CRM and analytics integrations. We pick the stack for the task — only what the team can maintain on its own.

Where do we start?+

With a free 30-minute call with the engineer: we work out how your team is set up and suggest a format that fits. You get an honest read either way: if we see value, we go on; if there is none, we will say so.

12Who you work with

Two people stand behind the practice: an engineer who builds and a practitioner who delivers

Tim Zinin
Tim Zinin
AI systems engineer

Builds AI tools for teams and clients. Writes the code personally. Puts AI on engineering rails — from a single agent to a company's whole AI infrastructure.

Anthropic · certified developer 27 founders trained
Alexander Minakov
Alexander Minakov
Product lead · AI transformation specialist

Over 20 years launching digital, AdTech and AI products. Founder of RTB-Media / Improvado, stake sold in 2017. Helps owners and management teams bring AI into processes, product, marketing and sales.

>$100M — Improvado's current valuation 20+ years in product and growth
13Applications

Q3 · 2026: applications open

We take a limited number of companies each quarter so every practice runs hands-on and at the team's pace. 3 of 8 slots are open — once they are gone, applications close until the next quarter.

Hold a slot — start with a call →
Q3 slots applications open
3 open of 8
5 slots taken · applications reviewed in order
3 slots open in Q3

Meet the engineer — the first call is online and free

Tell us how your team is set up. Tim will work out which processes AI can take over first and suggest a format. You get an honest read either way: if we see value, we go on; if there is none, we will say so.

Book a call →
▹ part of the Zinin AI network