Design. Teach. Test. Fly
Aircraft aerodynamic design • Flight tests • Machine learning • Open Science
Elena Karpovich — Associate Professor, PhD, Moscow Aviation Institute
Hi!




Here you’ll find everything I create and share: course materials for my students, research papers, code, datasets for machine learning models, and more.
Once a month, we also gather for the Journal Club — a friendly discussion of the latest advances in aircraft engineering.
Whether you’re passionate about theory, hands-on testing, or simply curious to learn something new in aircraft engineering — you are very welcome here.
I’m Elena Karpovich, Ph.D., Associate Professor and researcher at the Moscow Aviation Institute. Welcome to my open workspace!
Go to
I love helping design better aircraft — from aerodynamic shapes and wing sections to propellers and flight trajectory optimization. I use surrogate modeling and machine learning to speed up aircraft design, and I run flight tests on small UAVs.
Once a month, we discuss the latest in aircraft engineering. Everyone is welcome.
Learn more
Latest news
Hopefully, this website will bring much value to all guests:)
May 30, 2026
June 15, 2026
June 12, 2026
It has been my dream for so many months to write this code and apply it to a real project. Now, this dream came true.
Sensors, power source, wind-tunnel speed readings linked to the propeller control system, test program - checked and ready for a new adventure!
Now, we have to carefully postprocess the raw test data.
July 27, 2026
The code for the design and analysis of a variable-pitch propeller is ready to go!
Everything is ready for the wind-tunnel tests of the lift-cruise propeller.
This website is launched.
The article is about three lessons I learned while working hard on the variable-pitch prop design and analysis code.
A new blog article has been posted.


August 13, 2026
I launched a new small project:)



August 9, 2026
New ML models and a dataset are uploaded to the Codes page.



These are Random Forest Regression models for propeller mass, power, and efficiency prediction from required thrust trained on a BEM-based dataset.
I also created a page where I share the models' details, and where you can download both the models and the dataset.
I am truly happy to say: one more dream became a reality, and I can finally share it with everyone.
August 16, 2026
Made on
Tilda