Senior Edge AI Researcher – Hardware-Aware ML

Huawei Canada; Edmonton; On-site; CAD 90,000 - 140,000; Full time

Job summary

Huawei Canada is seeking a Researcher in the Software-Hardware System Optimization Lab to advance hardware-aware neural network optimization for consumer devices. The role focuses on quantization, mixed precision, pruning, and NAS, with emphasis on energy-efficient inference across ARM CPUs, mobile GPUs, DSPs, and NPUs.

You will prototype efficient methods, publish findings, and collaborate with compiler/runtime teams (TVM/MLIR/TensorRT). Mentorship and roadmap contributions are expected.

Qualifications

• PhD or equivalent research experience in ML/CS/EE.
• Strong coding in Python and C/C++ with AI tooling.
• Proven publication record at top ML venues and/or patents.
• 2+ years post-grad experience with impact in model compression or ML systems.
• Experience with edge hardware optimization (ARM NEON, mobile GPUs, NPUs).

Responsibilities

• Conduct research in hardware-aware neural network optimization (quantization, pruning, distillation).
• Develop latency/energy-aware training objectives and Pareto optimization.
• Prototype and evaluate efficient inference under device constraints.
• Publish findings internally and externally (papers, workshops, patents, blogs).
• Optimize inference pipelines across pre/post-processing, scheduling, and memory planning.
• Collaborate on compilers/runtimes (e.g., TVM, MLIR, XLA, TensorRT, ONNX Runtime, TFLite) to improve coverage.
• Profile and optimize models with real device traces and address bottlenecks (cache misses, bandwidth).
• Build hardware-aware benchmarking and regression suites for edge targets (ARM CPU, mobile GPU, DSP, NPU).
• Create deployment recipes for heterogeneous compute (CPU+GPU+NPU) including partitioning strategies.
• Drive on-device personalization and incremental updates when needed.
• Mentor junior researchers/engineers and raise measurement rigor.
• Define technical roadmap areas (next-gen quantization, kernel optimization, edge model families).

Skills

Python

C/C++

Research

Publications

Education

PhD or equivalent

Tools

PyTorch

TensorFlow

JAX

ONNX

TFLite

TensorRT

TVM

MLIR-based stacks

Job description

Huawei Canada is seeking a Researcher in the Software-Hardware System Optimization Lab to advance hardware-aware neural network optimization for consumer devices. The role focuses on quantization, mixed precision, pruning, and NAS, with emphasis on energy-efficient inference across ARM CPUs, mobile GPUs, DSPs, and NPUs.

You will prototype efficient methods, publish findings, and collaborate with compiler/runtime teams (TVM/MLIR/TensorRT). Mentorship and roadmap contributions are expected.