Not for everyone — and that is the point
Lemdranok runs focused masterclasses in AI for people who want to understand what is actually happening inside the models — not just click through a demo. Sessions run in groups of 12 and move at a pace that assumes you are paying attention.
What participants said after 6 weeks
Pavlo Hrytsenko
Data Analyst, Kyiv
Before this I had read 3 books on ML and still could not explain backpropagation clearly. After session 4 I wrote a 2-page explainer for my team and they actually understood it. The instructor stopped me mid-question twice because I was already answering my own problem — that felt good.
AI Foundations Intensive
Taras Bilyk
Product Manager, Lviv
I came in expecting theory. What I got was a session where we fed a real dataset into a classification model and watched it fail — then figured out why. That single exercise was worth more than 6 hours of video courses I had done before. My spec-writing for the dev team changed noticeably after week 3.
Prompt Engineering LabOlena Marchuk
UX Researcher, Kharkiv
I was skeptical that a 6-week course could do anything meaningful with something as broad as AI. I was wrong. The scope is deliberately narrow — we covered 4 core algorithms in depth rather than 20 algorithms at surface level. By week 5 I had built a working text classifier for user feedback tagging. It runs in production now, processing roughly 400 entries per day.
Computer Vision Workshop
A short list, not a catalogue
Three programs run at any given time. Each one covers a specific slice of AI — not a general survey. Group size is capped at 12 per cohort so that questions get real answers, not scripted ones. Sessions are live, recorded for review, and run in Ukrainian and English.
AI Foundations Intensive
8 weeks. Covers supervised learning, model evaluation, bias detection, and feature engineering using real tabular datasets. No prior ML experience required — basic Python helps.
Prompt Engineering Lab
4 weeks. Hands-on work with large language models — context windows, temperature settings, chain-of-thought prompting, and structured output generation. Practical from session 1.
Computer Vision Workshop
6 weeks. Image classification, object detection, and basic segmentation using annotated datasets. Participants leave with at least 1 working model they built themselves.
What becomes possible when you finish
Finishing a Lemdranok program does not hand you a certificate and a vague sense of accomplishment. It puts you at a point where specific tasks that were previously out of reach are now within reach. That is a narrower claim than most platforms make — and it is the honest one.
The average time between program completion and the first applied AI project — based on self-reported follow-up from 3 consecutive cohorts. Some participants move faster, some slower. Results depend entirely on what you do with the material.
Reading model outputs critically
You can look at a confusion matrix or a loss curve and know what it is telling you. You stop treating model results as a black box and start treating them as data to interpret.
Writing specs that engineers respect
Non-technical participants consistently report that their communication with ML engineers improves — not because they learned to code, but because they learned what questions are worth asking.
Running a small experiment independently
Every program includes at least 1 project you build from scratch. By the end, running a second experiment on your own — with a different dataset and a different goal — is a realistic next step, not a distant aspiration.
