Lead Edge AI Engineer
Core
Lead edge AI development for Grab's Geo Vision team, focusing on multi-task learning models and video action recognition systems for embedded Android devices to improve maps and road network intelligence.
Role type
Lead Edge AI Engineer (Computer Vision)
Builds
Production ML models and video action recognition systems deployed on Qualcomm Snapdragon edge devices
Domain
Transportation technology / Computer Vision / Edge AI
Deliverable
production ML models
Required skills
Multi-task learning model design, Video action recognition, PyTorch, Edge deployment on Android, Qualcomm SNPE/QNN SDK, DSP optimization, Model quantization and pruning, Resource-constrained environment management
Preferred skills
Dynamic graph execution on Hexagon DSP, Android hardware-software interaction, Sensor fusion (GPS/IMU/video), 3D reconstruction, Low-level system software on Qualcomm chipsets
Technologies
PyTorch, Qualcomm SNPE, QNN SDK, Android, Hexagon DSP
Responsibilities
Develop and refine multi-task learning models (Hydranet architecture), Deploy CV algorithms to embedded Android platforms, Conduct performance analysis to reduce power consumption and manage thermal constraints, Implement safety mechanisms for dynamic model graph reconfiguration, Collaborate with Firmware and Mobile teams
Seniority
Senior, hands-on IC with leadership responsibilities
