Senior Researcher - Edge AI Optimization/Hardware-Aware ML
Core
Conduct research in hardware-aware neural network optimization and efficient inference for edge AI, multimedia, and mobile gaming to enhance Huawei consumer device performance.
Role type
Senior Researcher (Edge AI/Hardware-Aware ML)
Builds
Efficient inference pipelines, compilers/runtimes, and deployment recipes for heterogeneous edge compute (CPU+GPU+NPU).
Domain
Edge AI, Mobile Systems, Hardware-Aware Machine Learning
Deliverable
production ML models
Required skills
Hardware-aware neural network optimization, Quantization-aware training, Mixed precision, Pruning, Distillation, Neural architecture search, Latency/energy-aware training objectives, Pareto optimization, Compiler optimization, Runtime optimization, Performance profiling, C/C++ programming, Python programming, Deep learning frameworks (PyTorch, TensorFlow, JAX), Edge hardware knowledge (ARM NEON, mobile GPUs, DSP, NPU), Model compression, Efficient architectures, Distributed benchmarking, Power/thermal constraint analysis, LLM/VLM inference optimization
Preferred skills
Experience with AI agent/toolchains, Experience with specific edge hardware architectures, Experience with efficient LLM/VLM inference techniques (KV-cache optimization, speculative decoding), Experience with CI for performance regression
Technologies
TVM, MLIR, XLA, TensorRT, ONNX Runtime, TFLite, ExecuTorch, PyTorch, TensorFlow, JAX, ONNX
Responsibilities
Conduct research in hardware-aware neural network optimization, Develop novel approaches for latency/energy-aware training objectives, Prototype and evaluate techniques for efficient inference under device constraints, Publish and present findings, Optimize inference pipelines, Collaborate on compilers/runtimes, Profile and optimize models with real device traces, Build and maintain hardware-aware benchmarking methodology, Create deployment recipes for heterogeneous compute, Drive optimization for on-device personalization, Partner with product engineering and hardware teams, Mentor junior researchers/engineers, Define technical roadmap areas
Seniority
Senior, hands-on IC