Why I built Bench

I've been teaching technical classes for a while now - chemistry at the University of San Francisco, then AI/ML at Fullstack Academy, then data science mentoring at 4Geeks Academy. The content changes. The problem doesn't: most online learning platforms optimize for passive consumption, not active work.

You watch a video. You maybe follow along. You close the tab and nothing sticks.

That's not how I learned to code, and it's not how I teach. When I was a graduate student running molecular dynamics simulations at the Texas Advanced Computing Center, I learned by doing - by writing the scripts, debugging the runs, reading the output. The bench, in my case, was a terminal.

The name

"Bench" is a short word with two meanings that both fit.

The first is the lab bench - the place where scientists do actual work. Not the classroom, not the lecture hall. The bench. That's where I spent my PhD years, pipettes in hand, running experiments I had designed and would have to defend.

The second is the workbench - a place where you build things. Same idea, different context.

I wanted a name that said something true about the experience. Bench says it.

The model

I didn't want to build another Udemy. I wanted to run live, hands-on sessions - the kind of class where you have a terminal open and you're working through something real, not watching someone else do it.

The practical problem with live classes is scheduling. If you put a date on the calendar before you know if anyone wants the class, you're gambling. If you wait until you have confirmed interest, you have no date to announce.

Bench solves this with a demand-driven queue. Students join an interest list for free. When enough people are interested, the class gets scheduled. No guessing, no empty rooms.

It's a simple idea, but it changes the dynamic completely. The schedule emerges from demand instead of being imposed on it.

What Bench is

Bench is a platform for individual, live, hands-on technical classes. One session. One GitHub repo. Real work.

The first class - How to Docker - is $10. It's a hands-on introduction to Docker for AI/ML. After that, classes will be priced at $20-50 per session.

If you've been meaning to learn Docker, or containerization, or anything else I end up teaching here - the interest list is free to join and there's no commitment until the class is scheduled.