Is there any good blog or video which describe how Uber/swiggy process timeseries gps locations of moving cabs/delivery agents and then how they assign cab/delivery agents for user request. For an overview I know that they can ingest geohash pings and then on the fly firing a geohash query for a particular radius but I want to have a in-depth understanding that how they are working at scale .
Swiggy has something called Just In Time assignments for most cases. There are other modes too. The driver pings goes to a system called the count-server. When an order is created, it checks the restaurant prep time and work backwards to assign the right rider at the right time. This is to avoid drivers unnecessarily waiting at restaurants, and cause no additional delays. Check out bytes.swiggy.com