As a Britisher learner it's frustrating¹ actually, because there is a standard for how to do this - IAST, for Sanskrit/derived generally - but of course that's not what native speakers use casually. E.g. your 'angrezi se hindi' would be 'añgrezī se hiñdī' but anyone writing Hindi with those accents is foreign or an academic. (Also people will casually write 'ay' instead of 'e' ए or 'ee' for 'ī' ई, etc. cf. 'paneer'.)
[1: The frustration is because it leads to ambiguity, whereas IAST is 1:1 and so preserves the phonetic nature of devanāgarī, and tells me exactly which t/d/r sound, if it's aspirated, etc. which a fluent native layperson's anglicised interpretation really doesn't. They might write gora & gora and know from context if that's gora or ghoṛā, but if I don't already know the word a gora like me is stuck.]
That should be 'aṅgrezī se hiṅdī' per IAST. In Devanagari: अंग्रेज़ी से हिंदी
If it were ñ instead of ṅ, the Devanagari would be अँग्रेज़ी से हिँदी which is incorrect.
The AI-specific use of em dash with the list of three near-synonyms following, the "it's not just this", the "it's a fundamental shift". This article seems AI-generated or at the very least AI-massaged.
Once you suspect AI has been used to write something, you're not sure if the 'author' has bothered to double-check the article for veracity, or if you're going to be doing this work for them.