They are responsible for (1) learning and improving a worldwide catalog of addresses with high-quality validation and geo-resolution, (2) building a places dataset to model where we delivery ranging from every single single-family home, campus, building, and apartment - along with their relationships and delivery critical attributes such as delivery hours, access information, mail rooms, delivery lockers, parking locations, entrances, and drop-off geocodes, (3) developing maps that capture a fresh and accurate road network, enable precise transit paths that optimize travel times while reducing travel risk in delivery routes and on-road navigation experiences and (4) developing feedback loops that leverage edge capabilities of millions of smart phones and tens of thousands of delivery vehicles to capture fresh street imagery, learn street signs, road markings, and road obstructions at scale, and reconstruct key delivery events and activities to improve the fidelity of address, place, and road datasets, optimize routes, and reduce defects. The Geospatial Science team is responsible for the quality and coverage of the core geospatial data, solvers, and real-time workflows that operate over petabytes of data, power trillions of transit time calculations daily, and operate on diverse environments spanning multi-modal cloud-based learning workflows, highly throughput and low-latency services, and edge compute applications on smart phones, delivery vehicles, and delivery stations.