Alexander Edelman
All work

Sber • 2026

Researching learning personalization before building it

How to personalize learning without building expensive infrastructure for nothing. I mapped the digital traces learners leave and proposed testing the hypotheses cheaply first.

The problem

The question was how to personalize learning without building expensive infrastructure that might turn out to be useless.

TODO: who the learners are and what made personalization a topic.

My role

What I did

TODO: what you did yourself.

Key decisions

Simplified and recreated for illustration. Details available in conversation.

Decision 01

Build the infrastructure or test first

  1. Build data infrastructure, then look for uses
  2. Test the hypotheses on mockups and a small sample firstChosen

Why

A hypothesis that fails on a mockup costs little. One that fails after the infrastructure is built costs the whole budget.

Test the riskiest assumption first.Simplified and recreated for illustration. Details available in conversation.

The artifact

Simplified and recreated for illustration. Details available in conversation.

Outcome

I presented the research to senior leadership, and they decided to test the hypotheses.

What I learned

TODO: 1–3 sentences.

The materials are internal and can’t be shared publicly. I’m happy to walk you through the details on a call.

More work

Sber • 2026A role-based compliance course on unfair market practices

The client needed proof that every employee had read the mandatory content. The course keeps that and builds practice around it, with situational tasks for each role.

  • Experiences
Read the case
Sber • 2026An AI-assisted design stage for course production

A research and design stage that runs before a course gets built. AI helps go from source materials, the brief and stakeholder interviews to a chain that starts with the behavior gap.

  • Systems & process
  • AI
Read the case