Multi-agent AI assistant at Magnit Tech
A multi-agent assistant for information retrieval, calculations, metrics and management reporting.
Explore project
Mark Serdyuk/Product Manager
From an idea
to a product
that works.
I’m Mark Serdyuk, a Product Manager.
I build AI products and B2B platforms.
Connecting research, technology
and measurable business impact.
100 M RUBeconomic impact / year
700+leaders use the AI assistant weekly
Currently at Magnit TechAI products & internal platforms
Experience & project work
Magnit TechYandexGazprom NeftT-BankX5 Group
01 Selected work
From a question
to a result.
Three projects that best
represent the way I work.
Launching “Mover without a van” at Yandex
“Mover without a van”: research, business case, user journey and production launch.
Explore project
AI automation research at Gazpromneft Business Service
CustDev across 10 subsidiaries, LLM applications and tools for evaluating automation impact.
Explore project
02 How I work
Understand the why.
Then build the how.
Understand
Needs, context, constraints. Interviews and data before choosing a solution.
Evaluate
Value, product metrics and economics. Clear criteria for success.
Launch
Priorities, requirements and teamwork. From a hypothesis to real use.
Improve
Feedback, impact measurement and the next product iterations.
My toolkit
CustDev · JTBD · CJM · Unit economics · SQL · LLM · Multi-agent systems · A/B tests
A small experiment / 01
What is
saved time
worth?
Move the sliders to explore the potential of AI automation for a team.
An illustrative model, not a calculation for any of my projects.
Automation potential
Annual value of time freed up4,8 M RUB
4,800 hours per year
Method and limitations
Users × minutes / 60 × hourly cost × 240 working days × adoption rate.
This estimates the value of time, not guaranteed budget savings. Implementation, model, support and quality assurance costs have not been deducted.
03 Now and along the way
Still
moving forward.
Follow via RSS ↗Current focus
AI products
at Magnit Tech.
Developing internal platforms, bringing AI into workflows and helping teams use it.
Moscow · September 2026
- Work
AI in leaders’ everyday work
At Magnit Tech, launched an AI assistant with 700+ weekly users and trained over 150 product managers and team leads in AI tools.
- Launch
A new service scenario at Yandex
Brought “Mover without a van” from needs research into production.
- Teaching
Sharing what you know
Developed business analytics materials, reviewed 100+ cases and taught classes of up to 170 students.
04 The person behind the work
Curiosity is
a skill, too.
I’m Mark. I enjoy making sense of complex problems, turning that understanding into working products and sharing what I learn.
Beyond product work: teaching, piano, sport and generative art. I created the CU:/dev/art exhibition and founded a student sports club.
Alongside my product career, I explore entrepreneurship through a GenAI project in the beauty industry.
2024—2028
Central University
Mathematics and Computer Science · Product Management.
2025—2026
Teaching business analytics
100+ cases reviewed, classes for audiences of up to 170 students.
2025
Gazprom Neft Corporate University
A programme for project leads and product teams.
scholarships
Potanin Foundation, T-Bank, Central University, Avito, Sber, Reksoft Consulting, Moscow Exchange.
hackathon wins and podium finishes
AI&Business SPB, “Idea. Code. Release.”, Datathon 2.0.
DEADLINE finalist and mentor
From the 2024 final to mentoring podium teams in 2025–2026.
More projects and research
T-Bank
131 in-depth interviews with auto services, a partner pipeline for 5 regions and an acquisition landing page.
X5 Group
Field research across 9 regions, a CJM and 50+ proposals for the partner directors’ authority system.
National Education Strategy 2040
Led a working group of 8, gathered input from 100+ experts and presented the results to government bodies.
Research and BRICS
Research on retail investors in BRICS countries and a payment transformation proposal, presented in Brazil in May 2025.
05 At a glance
Questions and answers
What does Mark Serdyuk do?
I am a Product Manager working on AI products, B2B platforms and internal services. I currently work at Magnit Tech; previously, I launched a service scenario at Yandex and researched automation at Gazpromneft Business Service.
What is my experience with AI adoption?
At Magnit Tech, I built and launched a multi-agent AI assistant with over 700 weekly users in leadership roles. I also trained more than 150 product managers and team leads in AI tools. Read about the project.
Where can I find projects and my CV?
The site presents three projects with their challenges, my role and outcomes. My full experience is available in the Russian-language CV. Projects · CV PDF · RU ↗
How can you contact me and follow updates?
You can contact me on Telegram to discuss a product role or an AI challenge. I share milestones in the Now section, with an RSS feed for updates. Telegram · RSS
The next conversation
Good products
start with a
conversation.
A product role, an AI challenge, or simply a catch-up.
✳ PM Runner A little break between sprints
Jump over bugs and deadlines. Space, ↑ or tap to jump. Double jump is available. Coffee adds 10 points.
Score: 0Best: 0Speed: normal