Building Interactive Worlds Through Physics and Design
We're a small group in Ulaanbaatar who spend our days thinking about physics engines, collision detection, and how to make digital objects feel like they have actual weight. Started in 2019, we've learned most of what we know by breaking things and figuring out why.
Our approach focuses on understanding Newtonian mechanics first, then translating that into code. We've spent years working with rigid body dynamics and soft body simulations. Teaching others means showing them the math behind momentum transfer and explaining why certain approximations work better in real-time environments.
Building believable spaces requires thinking about how light bounces, how surfaces interact, and what makes a room feel right. We work with students on spatial composition and teach them to spot when something feels off in a virtual environment. It's about observation skills as much as technical knowledge.
Making things respond properly to player input takes patience and iteration. We cover collision detection algorithms, response calculations, and the debugging process that comes with any physics-based interaction. Students learn through hands-on projects that usually don't work correctly the first dozen times.
Two people who've been doing this long enough to know what works and what's a waste of time. We're still learning new things every month, which is what keeps this interesting.
Lead Physics Instructor
Worked as a simulation engineer for six years before switching to education in 2021. Spent too much time optimizing collision systems for mobile devices. Now helps students understand why their physics calculations are eating up frame rates.
3D Systems Instructor
Started as a level designer, moved into technical art, eventually landed in education. Good at explaining why things look wrong in 3D space and how to fix them. Believes most problems come from not understanding coordinate systems properly.
Most game development courses focus on tools and workflows. We focus on understanding why things work the way they do. If you know the underlying principles, you can adapt to any engine or framework that comes along.
Our students work on projects that actually break. They debug collision systems that fail spectacularly and learn to read stack traces without panicking. We think making mistakes in a controlled environment is the fastest way to build real competence.
Classes run from September 2025 through March 2026, meeting twice weekly. We keep groups small because individualized feedback matters more than lecture content. Most learning happens during office hours when students are stuck on specific problems.