@itnoles are you referring to
AMPER-4366? We didn't have time to look into it yet.
At the moment, Amper is highly parallel and builds and tests all modules concurrently, launching a bunch of child processes (such as the Kotlin native compiler, or the JUnit test JVM). There is a very high chance that the memory usage comes from this rather than individual things consuming lots of memory. But again, we haven't investigated yet. We'll get to it, for sure!