As a native speaker of an Indo-Aryan language (Hindi), I would like to add some inputs.

Early Old Indo-Aryan (Vedic Saṁskr̥ta-Bhāṣā) and Later Old Indo-Aryan (Classical Saṁskr̥ta-Bhāṣā) had ā-ending masculine words: Sakhā (male friend; root word in itself), Pitā (nominative of Pitr̥ (father)), Ātmā (nominative of Atman (self)), Karmā (nominative of Karman (actioner)), Dharmā (nominative of Dharman (righteousness wielder)), Varmā (nominative of Varman (warrior)), Śarmā (nominative of Śarman (good)), Savitā (nominative of Savitr̥ (sun)) etc.

In Hindi (and I assume some other modern IA languages), some ā-ending masculine words ended up changing gender to become female: main examples I can think of being Ātmā and Savitā. Conversely, Pitā, Karmā, Dharmā, Varmā and Śarmā retain masculine gender.

Old Indo-Aryan also had plenty of masculine and feminine words ending in ī (the long vowel, as in English need): Tējasvī (male)/ Tējasvinī (female), Ōjasvī (male) / Ōjasvinī (female), Yaśasvī (male) / Yaśasvinī (female) Sarasvatī (female), Nadī (river; female) etc. But a similar gender transformation from masculine to feminine from OIA to Hindi happened for the word Sākṣī (witness).

What I am trying to convey is that ending vowel doesn’t necessarily indicate specific gender of an animate object in Indo-Aryan languages — there are plenty of words that break the rule of ā-ending (or other sounds) words being female.

Thanks.