模型优化算法专家-PICO
Core
Design and implement end-cloud collaborative content compression and transmission schemes for XR glasses, optimizing VLM integration for interaction efficiency and image compression algorithms for edge constraints.
Role type
Senior IC machine-learning engineer (XR compression & VLM optimization)
Builds
End-cloud transmission pipelines, VLM-based interaction systems, and XR image compression engines
Domain
XR (AR/VR/MR), Edge Computing, Computer Vision
Deliverable
production ML models
Required skills
C/C++/Python, VLM application & optimization, model deployment & quantization, token pruning, end-cloud joint optimization, edge inference & memory optimization, hardware acceleration (GPU/NPU/DSP), network adaptation
Preferred skills
LLM/Agent fundamentals, top-tier conference publications, AR/VR/MR product experience, deep learning-based compression (super-resolution/denoising), reinforcement learning (PPO/GRPO)
Responsibilities
Design end-cloud compression schemes under low-latency/power constraints; Integrate VLM for semantic understanding and bitrate allocation; Research and implement image compression algorithms (prediction/entropy coding/RDO); Build evaluation metrics and data loops for AR/XR scenarios
Seniority
Senior, hands-on IC
