Transport & on-demand
RidePulse
Geospatial matching and trustworthy ETAs
A ride-sharing backend answering "who is nearby" in milliseconds, with ETAs computed against live routing data rather than straight-line guesses.
This example comes from the technical delivery experience behind AussieSync. It does not necessarily represent a project contracted directly through AussieSync.
Have a process like this one? Tell us about it.
01 Problem
Matching a rider to a driver requires answering a geospatial question in milliseconds, against locations that never stop moving.
- Scanning every driver to find nearby ones does not survive a growing fleet.
- Straight-line distance ignores rivers, one-way streets and traffic, so ETAs were wrong.
- Riders abandon a booking when the arrival estimate is visibly untrustworthy.
- Location updates arrive continuously from every active device at once.
02 The solution
Geospatial indexing to answer proximity in milliseconds, with a matching engine that weighs more than distance.
- Driver positions are held in a geospatial index, so proximity is a lookup, not a scan.
- Matching balances proximity, direction of travel and rating rather than distance alone.
- ETAs are computed against live routing data, so they reflect the road network.
- Location and trip state stream over persistent connections instead of polling.
03 Technical approach
Every choice earns its place.
- Redis GEO & PostGISGeospatial indexes that make "who is nearby" a millisecond lookup.
- Go & Node.jsMatching engine and trip lifecycle services.
- WebSocketStreams live location and trip state to both sides of the ride.
- Routing APILive road-network data behind every ETA.
04 Outcome
Sub-second nearby-driver lookup, and ETAs riders can trust.
- Sub-secondnearby-driver lookup as the fleet grows
- TrustworthyETAs computed on live routing data
- Highertrip completion rates