Leading large scale logsim agent simulation.

Agent simulation

• Gained hands-on experience in almost all aspects of agent simulation, including perception sanitization, decision to handover to sim agent, intersection, trajectory proposal, refinement and selection, and the development of distributional metrics and other evaluation metrics, with varying levels of expertise in each area.

Network Infra Team (simulation)

• Designed and implemented a Go library based on shortest path algorithms that can simulate network traffic between B2 and data center in Google network

• Integrated this library to existing network impact analysis software used in Google and tested it with real demand data in Google network