In this paper, we explore the case of an aggregator-cum-restaurant that also offers pickup and delivery services to third party restaurants registered with it. The aggregator must decide on its fleet size and the optimal routes to assign to each vehicle deployed. We propose a heterogeneous, compartmentalised vehicle routing model with pickup and delivery for the aggregator involving time windows and source selection, to minimise the route duration (or the total cost) of its fleet. The model accounts for traffic conditions (captured by speed data) over the route, maximum service radius of the fleet and time windows for customers as well as restaurants. This paper, to the best of our knowledge, is probably the first one that deals with vehicle routing problem for an online hyperlocal food service marketplace (also referred to as aggregator) that functions as a quick service restaurant (QSR) as well.