Embedded ML Engineer for AR/VR Real-Time Inference

BCforward; Redmond; Vor Ort; AUD 172.000 - 258.000; Vollzeit

Zusammenfassung

BCforward is seeking a contract ML Software Engineer to bring state-of-the-art machine learning from lab to device. You will train models with researchers and own optimization and porting to embedded targets, balancing latency, memory, and power on constrained hardware.

The role emphasizes real-time inference, cross-functional collaboration, and building infrastructure to accelerate prototyping for AR/VR wearables.

Qualifikationen

• BS degree or equivalent work experience in a technical field.
• Experience optimizing and running ML inference on embedded/resource-constrained devices.
• Experience training ML models using frameworks such as PyTorch or TensorFlow.
• 2+ years of industry experience writing C/C++.

Aufgaben

• Collaborate with researchers to move ML models from prototype to real-time inference on embedded devices.
• Train, fine-tune, and adapt ML models considering latency, memory, power, and compute constraints.
• Optimize models for on-device inference using quantization, pruning, distillation, and operator fusion.
• Port and profile ML models on embedded targets (DSPs, NPUs, microcontrollers, mobile SoCs) and resolve performance bottlenecks.
• Write clean, high-performance C/C++ for on-device inference pipelines and integrate into prototype systems.
• Build tools to accelerate evaluation of model performance and accuracy on target hardware.
• Coordinate with hardware, runtime, and toolchain teams to ensure smooth integration.
• Work as part of a multifuntional team emphasizing communication and documentation.

Kenntnisse

Optimizing and running ML inference on

Training ML models

C/C++ development

Ausbildung

BS degree in CS/CE/EE or related field

Tools

PyTorch

TensorFlow

ExecuTorch

TFLite/LiteRT

ONNX Runtime

Jobbeschreibung

BCforward is seeking a contract ML Software Engineer to bring state-of-the-art machine learning from lab to device. You will train models with researchers and own optimization and porting to embedded targets, balancing latency, memory, and power on constrained hardware.

The role emphasizes real-time inference, cross-functional collaboration, and building infrastructure to accelerate prototyping for AR/VR wearables.