Daniil Evstratov — Senior Instructional Designer · LXD · Team Lead
Education
that makes
money.
I design courses — and the AI pipelines that produce them. 85+ courses shipped, a product serving 500,000 monthly users, and over $12M in revenue from learning products I built.
Looking for: Senior Instructional Designer / Learning Team Lead · remote · full-time or project
A course took a month. Now — a day. ×26
Every ruble in — thirty back. ROI 2982%
My most profitable course? I didn't write it. I rebuilt it
Users aren't dumb. Your course is boring. Retention ×3
4 launches in 3 months. $350K+ revenue
My first learning product was people
Sources: Eduson (B2C/B2B EdTech, 2023–2025) and Tixu.ai (AI education, USA/EU, 2025–2026), internal company analytics. Happy to walk through the methodology in an interview. Revenue converted from RUB.
Case studies.
0 → 500,000 MAU as the only instructional designer
Context
Tixu.ai — an AI education product for the US/EU markets. All learning content was one person's responsibility: mine.
What I did
Designed 85+ courses and 1,100+ lessons in English. Worked with the CPO and two PMs on product hypotheses and metrics. Built the production system so one designer could serve a product of half a million users.
Result
From zero to 500,000 MAU in a year. Course CSI — 80%+.
×26: a multi-agent course factory — 20+ AI agents in one pipeline
Context
The product was growing faster than a human could write lessons. Growing the team is slow and expensive. I solved it with architecture: not "chatting with an LLM," but a system of agents with roles, rules and quality control.
What I did
Designed and built a multi-agent system: an orchestrator routes tasks, program and lesson pipelines push drafts through an "analyze → rewrite → checks → auto-repair" loop, and a separate agent layer localizes content into 4 languages. Quality is enforced by deterministic linter scripts and two human gates — not eyeballing. The token guardrails and batching aren't theoretical: an early version of the pipeline burned $2,000 in two weeks, against a previous peak of $550 a month. It doesn't anymore.
Result
Shipping one course became 26× faster: one learning designer produces the output of a whole team. 20+ agents in daily use, localization EN → ES / FR / DE / JA, −50% token cost.
How the pipeline works
→ the diagram scrolls horizontally
ROI 2982%: the record-setting first month
Context
Eduson, B2C/B2B courses. I led the course end-to-end as its tech lead — from methodology to release.
What I did
Built an AI course in two weeks with minimal resources — full cycle, from curriculum to launch. It answered the market's most urgent demand and lifted the company's whole catalog.
Result
2982% ROI in the first month after release.
$1.2M a year — from a course rebuilt in 2 weeks
Context
The business was sitting on a pile of ready-made content. A new course from scratch takes months; the market window was weeks.
What I did
Ran deep jobs-to-be-done research, mapped the methodology and skill tree — and rebuilt the existing content into a new program in two weeks.
Result
Over $1.2M in revenue in the first year.
Retention ×3: redesigning the learner journey
Context
Users were churning in the first days. The default answer is "make more content". I went to the data instead.
What I did
Focused on the first-time user experience: designed experiments with the PMs, ran them, and reworked the learning logic of the first lessons.
Result
Day-7 retention tripled.
The team: 4 launches in 3 months and designers who grew
Context
Eduson, lead of the instructional design team. Led two teams in parallel and owned the project triangle: budget, deadlines, resources.
What I did
Built processes and quality standards, ran the department's internship, hired 4 people full-time, grew one specialist into a team lead. Refactored three products.
Result
4 launches in 3 months with $350K+ net revenue. The team over-delivered for six months straight. CSAT 91%+, CSI 94%+.
Russia's #1 counselor school: a methodology that scales through people
Context
Before EdTech I taught psychology and pedagogy at a university — and was named teacher of the year. In parallel, I coached teachers at a counselor school.
What I did
Ran workshops for teachers before and during the course — from behavioral psychology to public speaking. Facilitated classes and gave feedback after every lecture.
Result
The school ranked #1 in Russia. And by the end of the course, students started showing up dressed as me — the most honest engagement metric I've ever seen.
How I work.
Methodology
I design the learning system as a whole: the learner journey, measurable outcomes, engagement that carries people to the finish. Andragogy and instructional design are my daily toolkit.
Product
Every course has unit economics, retention and CSI. I count the money and validate hypotheses before the first lesson is written.
AI production
I build LLM pipelines: generation, critique, editing, localization — with quality gates at every step. One designer ships the output of a whole team.
Process, backed by data
- Every working hour is tracked. Task estimates come from my own statistics, not gut feeling.
- Each type of design task has a workload coefficient — quarterly plans match reality.
- Teams get weekly decomposition and a transparent roadmap: deadlines are never a surprise.
What colleagues say.
videos in Russian, subtitled — click to watch with sound
"Delivers beyond what's asked. A reliable person you can truly count on"
"The team over-delivered on course production plans. I always knew Daniil would protect the team"
"Someone who truly hears the business context — and keeps content production firmly in hand"
"The courses he built — like the one on neural networks — were warmly received and sold actively"
"I completed my internship thanks to his leadership and joined the team full-time"