Help please

[Unicode-Hater]

I saw that you created a language model and I want to know what ate the steps in order to create one?
More context:
I'm member of the a robotics team and we are coding a speech recognizer for demonstrative purposes, our teacher wanted to be exclusively offline and we found out Mozilla DeepSpeech, which is not going to work because we have to build it on a Raspberry Pi (preferably 0) because of ram, cpu limitation and the fact that Raspbian is 32 bit. Yes I know that there is a documentation on the cmusphinx site, but we (and exclusively I) are not experienced enough to do it and yes, I'm conscious about the recomandation of those at cmusphinx to not try it unless experienced. If you are curious what language are we trying to create a model, is Romanian, but we want to do a small model with 50-100 words that it should know.