My teaching

My teaching, over the years

I’ve taught in a few formats:

• “Advanced Machine Learning” tutorials in a master’s program
• C programming in university courses
• A full “Deep Learning” week at Le Wagon
• A three-day course on “Machine Learning in Healthcare,” followed by a week-long hands-on project track, at the Lviv Data Science Summer School

I still occasionally teach in AI training programs.

Teaching has a superpower: it tests what you truly understand

Teaching is the fastest way to find what you don’t actually understand. You feel it the minute students start asking questions.

If you can’t explain something clearly, you don’t own it yet.

That’s why I think everyone should teach.

Not necessarily in a classroom. Explain your work to someone else. Write a tutorial. Mentor a junior. Give a talk to non-specialists.

Teaching forces clarity.

Math is the only field people are proud to be terrible at

While teaching, I noticed something strange: people are proud of being bad at math.

You rarely hear the same tone in other domains. But “I was so bad at math” is often said with a mix of relief and pride.

That probably has less to do with math itself than with what we’ve turned it into—and how we teach it. Too often, we start with abstract objects that feel useless: derivatives, integrals, matrices. Students memorize procedures. Exams become memory tests.

It doesn’t have to be that way. Many ideas can be introduced simply.

I often use functions as an example. In high school, we start with linear functions—ax + b—then quickly lose the thread. Even Wikipedia can make it sound intimidating: “a linear function from the real numbers to the real numbers is a function whose graph (in Cartesian coordinates) is a non-vertical line in the plane.”

There’s a much easier entry point. Show someone an x–y graph. Put age on the x-axis, from 20 to their current age. Ask them to draw their salary on the y-axis.

They’ll sketch something non-linear, rarely monotonic, full of jumps, plateaus, and reversals—sometimes even discontinuities. In other words: a function that actually looks like life.

Real-world examples can introduce complex concepts—and make them easier to formalize.

More broadly, I wish we taught more logic: how to reason, structure an argument, spot contradictions, and defend yourself against bad reasoning. Programming helps, but logic is bigger than code. It’s mental hygiene.