Vlad Popov.

Dr Vlad Popov · Product leader, EdTech and AI in education

Product decisions in education, tested with real users before the big spend.

Product leader in international student recruitment and learning. Twelve years as Chief Product Officer of StudyQA, a study search platform used by 4.3 million students a year. PhD in higher education policy from the University of York. I prototype the riskiest part of an idea myself, measure it, and hand a proven brief to the team.

See the work Download CV
  • 4.3Mstudents a year on StudyQA
  • 10,000+universities in the catalogue
  • 12 yrsas CPO, team of 20

Selected work

Each item shows the problem, the decision I made and what happened. Education and admissions first, then data products and smaller market tests.

StudyQA home page with the Academic Program Finder search

EdTech · International student recruitment · 2014 to present

StudyQAStudy search platform that helps students compare programmes abroad and sends qualified enquiries to universities.

Problem. Students choosing a degree abroad face thousands of options and no neutral place to compare them; universities pay for enquiries that are often unqualified.

Decision. Own the full journey in one product: search, enquiry routing, CRM and partner reporting, so both sides see what works. Modernise the platform in place instead of a risky rebuild.

My role. Chief Product Officer: strategy, roadmap, cross functional team of 20, partnerships with British Council, DAAD, QS and American Councils.

  • 4.3Mstudents a year
  • 10,000+universities, 46,000+ programmes, 100+ countries
  • 86automated workflows across content, communications and reporting
  • 0downtime across 4 major framework migrations

Stack: Laravel, Vue.js, MySQL, CI/CD, 7 servers

studyqa.com ↗
Sample admissions readiness report for HSE University: overall score 70 out of 100 with scores by area
Report in Russian

Admissions · B2B · March to May 2026

Admissions readiness auditRussian university websites, scored the way an applicant sees them.

Problem. Universities pay to promote their programmes, but rarely see how their own websites lose applicants on the way to an application.

Decision. Open the conversation with evidence: a free, personal report for each university with a score, peer comparison and concrete fixes. The audit email drew 93 clicks, the best of five campaign waves.

My role. Defined the scoring model and the offer, built the scanner and report pages, ran the campaign.

  • 417university sites audited, one report page each
  • 50/100average score across 168 scored universities
  • 21.1%open rate on 5,495 delivered emails, first wave
  • 183target actions in a paid campaign for a university client that followed
  • Audited sites without
  • Admission chance calculator99%
  • Structured data for search92%
  • Chat assistant for applicants82%

Stack: PHP scanner, Node.js report generator, Laravel CRM, NotiSend

Sample report ↗
AI University by Stadika: catalogue of AI courses translated into Russian, with 65 courses and 1,600+ lessons
Interface in Russian

Learning product · Russian market · 2026

Stadika AI UniversityPaid learning platform with AI courses in Russian, for professionals and students.

Problem. Good AI courses exist only in English, and many learners in Russia cannot use them comfortably.

Decision. Localise open course material from AI providers instead of writing a catalogue from scratch; free first lessons, one payment for full access, attribution on every order.

My role. Product owner and hands on builder: catalogue, paywall, payments, Telegram onboarding, analytics.

  • 65courses, 1,600+ lessons in Russian
  • 491lesson videos dubbed into Russian, 36.8 hours
  • 144paid orders, May to September 2026

Stack: Vue.js, PHP, Telegram bot, Yandex Metrica

university.stadika.ru ↗
Pumpkin Smash game start screen on a phone, with scoring rules and a Play button
Interface in Russian

Growth · Gamified acquisition · October 2026

Pumpkin SmashA seasonal game for Stadika and its AI University, on the web and as a Telegram Mini App.

Problem. The AI University needed new learners and a cheap way to test whether a personal discount converts, without a large ad budget.

Decision. Reward what the business needs (channel subscriptions, referrals), give every player a personal promo code that works at checkout, and set the prize budget to break even after three paid orders.

My role. Game and funnel design, rules, build, analytics.

  • 5tracked funnel events, from game end to checkout click
  • 3paid orders to break even
  • 84automated tests on the game API
Launches 20 October 2026, results to follow

Stack: PHP, MySQL, Telegram Mini App, Yandex Metrica

Analytics · Marketing attribution · 2026

Channel attribution dashboardOne view of ads, web analytics, enquiries and payments for Stadika and its AI University.

Problem. The team was deciding where to spend on ads without knowing which channel actually brought paying customers.

Decision. Fix the data before buying more traffic: first touch attribution for 90 days, daily sync of ad spend, analytics goals and payments, aggregates only with no personal data on screen.

My role. Defined the metrics and attribution rules, built the dashboard inside the admin panel.

  • 1,115consultation requests found that had not reached the CRM since January
  • 666enquiries and 37 payments since July in one view
  • 7dashboard tabs for the marketing team

Stack: Laravel, Filament, Yandex Direct and Metrica APIs. Dashboard is internal, so no screenshot.

UK Company Data page on globalmindsetmarketing.co.uk with status prototype, early access

Data product · UK · 2024 to present

UK Company DataStructured financial data from UK company accounts filed at Companies House, for analysts, research teams and AI agents.

Problem. Many UK accounts arrive only as scanned PDFs with no machine readable tags, so private companies are hard to compare at scale.

Decision. Test accuracy on a measured pilot before scaling, and link every value to its source filing so users can check it.

My role. CTO of Global Mindset Marketing Ltd. Status: prototype, early access, pre revenue.

  • 800,000+iXBRL filings parsed into 27 standard fields
  • 3.5M+PDF accounts processed on Isambard-AI, 20,000 GPU hour UKRI AIRR allocation
  • 88.8%field accuracy on the pilot set; 86.6% on a 2,656 document test set

Stack: Python, vision language extraction, MCP server

Project page ↗
IA con Clase home page: practical AI courses in Spanish for Latin America and Spain
Interface in Spanish

Learning product · Market entry test · October 2026

IA con ClasePractical AI courses in Spanish for Latin America and Spain, starting with Mexico.

Problem. Would a proven Russian course model work in a new language and market? Guessing would be expensive.

Decision. Launch with a small paid search budget and a stop rule agreed before the first click: enough clicks and no checkout means pause and rethink.

  • 37courses and 996 lessons live in Spanish
  • 62clicks, 0 checkouts in stage one: paused by the rule
  • £33spent to learn it; stage two narrowed to one offer

Stack: Cloudflare Pages and Workers, Stripe, Google Ads

iaconclase.com ↗
Claude Habitat desktop app showing the personal agents it set up for a user

AI workspace · Open source · April 2026

Claude Habitat and TeleClaudeAI workspaces for non technical teams, set up through a short conversation instead of configuration.

Problem. Small teams want AI assistants for their real work, but setup needs a terminal and a developer.

Decision. Make onboarding an interview, and give each project its own Telegram topic with its own memory, the way people already work.

  • FinalistAnthropic hackathon, top 2.4% of about 20,000 applicants
  • 107project topics routed through TeleClaude

Stack: Electron, TypeScript, Telegram Bot API, LanceDB

GitHub ↗
TokenToad landing page with a pixel toad mascot

Browser extension · 2026

TokenToad

Teams find out what they spent on AI APIs at the end of the month. TokenToad shows spend and subscription usage in the browser toolbar, with budget alerts before the limit.

Two modes: API billing and subscription limits. Status: landing page live, Chrome Web Store release pending.

Stack: Chrome extension (Manifest V3), Preact, Vite

SimpleMTD landing page for UK landlords with a waitlist and pre order button

Smoke test · UK landlords · May 2026

SimpleMTD

Making Tax Digital for Income Tax is new for UK landlords. I tested demand with a landing page, a waitlist and a pre order price before building the product.

Result: one application. Not enough to justify a build, so it did not go beyond the test.

Stack: Cloudflare Workers, Stripe, GA4

How I work

I build prototypes to reduce risk, not to replace engineering. The team gets a tested idea and real numbers, not a slide.

  1. Discover

    Talk to students, universities and staff; read the funnel data to find where people drop off.

  2. Prototype

    Build the smallest working version that tests the riskiest assumption with real users.

  3. Measure

    Agree the success metric and the stop rule before launch, then report what actually happened.

  4. Hand over

    Give the team a proven brief, the data and the open questions, then support adoption and training.

Writing and speaking

Research on how universities respond to markets, and practical work with the people who run admissions.

  • Book · 2024

    Degrees of Commerce

    Navigating the Marketization of Higher Education. History, theory and world models of university marketisation, and how it shapes academic integrity, access and equity.

    304 pages · ISBN 979-8876523433 · Amazon UK · Amazon US

  • PhD · University of York · 2022

    Universities under market pressure

    Doctoral research on how universities respond to market pressure and governance reform. The lens I still use when I design products for admissions teams.

  • Teaching and events

    Digital recruitment seminars

    Project 5-100 seminars on digital recruitment for faculty, deans and university leaders. Participant of the St Petersburg International Economic Forum on EdTech, AI and strategic partnerships.

Dr Vlad Popov

About

I have led product at StudyQA since 2014, working with university recruitment teams across several countries and with partners such as British Council, DAAD, QS and American Councils.

My PhD at the University of York (2022) studied how universities respond to market pressure and governance reform. I wrote Degrees of Commerce: Navigating the Marketization of Higher Education (2024) and ran Project 5-100 seminars on digital recruitment for faculty, deans and university leaders.

I am looking for a senior product role in UK EdTech or higher education: Chief Product Officer, Head or Director of Product. Based in Liverpool, open to London and hybrid work, UK Settled Status.