I suspect this is a form of code switching. When people are talking in their second language, and hit upon a concept that leaves them lost for words, they drop back on their first language to vocabulary to fill in the gap.
For example, I was having a conversation with a Bulgarian yesterday about his diet and he said, “I like to eat this thing with grilled meat on a stick, kebob, its a long stick with meat on it. I would eat kebob very often.”
This Bulgarian didn’t know the English word for this food was kebab, so he drops into his native language and accent for that word.
Of course, for this Bulgarian , he was speaking not writing. Had he been writing, he would have used the Latin script to write kebob. Bulgarian has its own script but transliterating from Bulgarian script to Latin script is relatively simple.
Also note that once the first language word has been introduced into the conversation, the speaker will continue to use it liberally, without further definition, assuming the audience has learned that word.
Now imagine the same conversation but with a Chinese speaker: “I like to eat this thing with grilled meat on a stick, yang rou chan, its a long stick with meat on it. I would eat yang rou chan very often.”
In this case, I’ve used pinyin to transliterate the Chinese words, however, pinyin isn’t intuitive to many Chinese writers. They can write in Latin script for writing English, but when thinking in Chinese, their native Chinese script dominates for writing. They might give you the pinyin as an afterthought, or they might not.
To their mind, the pinyin isn’t the word, they have to use their native script to convey their native language. So you might get something like: “I like to eat this thing with grilled meat on a stick, yang rou chan (羊肉串), its a long stick with meat on it. I would eat 羊肉串 very often.”