Over the past weeks, RideLvL has moved from concept to real-world testing.
We’ve now locked our initial tech stack around React Native and built a first MVP designed specifically for instructors. This early version is already being tested with ski and snowboard instructors in France and Switzerland, across the Alps and the Pyrenees.
From day one, our goal has been simple: get out of the doc and onto the terrain as fast as possible.
Why instructors matter so much to us
Instructors are central to RideLvL.
They’re the ones who:
understand progression in real conditions,
interact with riders every day,
and feel immediately when a tool fits the flow of a lesson… or doesn’t.
Working closely with them allows us to confront our assumptions quickly and iterate fast, before building anything too far from reality.
This has been a true win-win so far.
Many instructors already use video with their students. Some film short clips and share them via WhatsApp or messaging apps. The intent is right, but the tools aren’t designed for coaching. Videos get lost in threads, feedback isn’t structured, and nothing really carries over from one session to the next.
The feedback we’ve received has been consistently encouraging. Not because the MVP is perfect (it’s not), but because the problem resonates immediately.
One constraint above all others: simplicity
One thing has become very clear.
If using a digital tool adds work for instructors, it won’t stick.
Lessons are dense. Attention is split between safety, technique, terrain, and group dynamics. Anything that requires extra steps after the session quickly becomes a no-go.
That’s why we’re designing RideLvL to fit as much as possible during the lesson itself. One direction we’re actively exploring is speech-to-text, allowing instructors to comment on a video naturally, without stopping to type or switch context.
One area we’re actively exploring is speech-to-text as instructors film and comment during a lesson. When an instructor speaks while recording, their feedback can be automatically matched with the right moments in the video, then structured into clear, professional takeaways for the rider.
The bar is high: if it doesn’t feel lighter than WhatsApp, it’s not good enough.
How AI fits into this approach
For this first MVP, we deliberately rely on existing AI building blocks where they already work well, such as speech-to-text, pose estimation, or language models to help structure feedback. This allows us to move fast on the terrain and focus our energy on what really matters: how these tools fit into real lessons, real gestures, and real coaching moments.
In parallel, we’re developing our own models and algorithms designed specifically for action sports coaching. As instructors use RideLvL in real lessons, their annotations and feedback help us validate, refine, and train these features, so they genuinely augment coaching and serve riders, without adding complexity.
Who we’re focusing on first
At this stage, our primary focus is on:
independent instructors,
private ski schools,
and one-on-one lessons.
These contexts allow for faster experimentation and tighter feedback loops. We’re also starting to explore potential tests with training clubs, where progression tracking over time becomes even more relevant.
Staying close to the terrain
This phase is deliberately hands-on.
We’re continuing to onboard instructors, test the MVP in real lessons, and refine the product based on what actually happens on snow.
If you’re a ski or snowboard instructor, or if you work closely with instructors and this resonates, we’d love to stay in touch.
You can reach out directly, or check the page dedicated to instructors here:
With the season now well underway, we’re continuing to test and refine RideLvL in real lessons. We’ll share what’s worth sharing as things evolve on real terrain. Ride on.
