Run parallel api inference for QA

[Mamoon]

How can we handle the parallel request of question-answering using API inference for qa?