onsite
Audio Machine Learning Engineer - Qualcomm
ML Engineer
Develop, optimize, and deploy real‑time audio machine‑learning models on Snapdragon CPUs, GPUs, and NPUs, ensuring low latency, minimal memory usage, and power efficiency for embedded audio systems.
About the role
Key Responsibilities
- Design and train state‑of‑the‑art audio ML models using frameworks such as TensorFlow and PyTorch.
- Optimize models for real‑time inference on Snapdragon CPUs, GPUs, and NPUs, applying quantization, pruning, and hardware‑aware tuning.
- Convert and integrate models into ONNX or other portable formats for seamless deployment on embedded platforms.
- Collaborate with firmware and DSP teams to meet strict latency, memory, and power budgets.
- Validate model performance on target hardware, conduct profiling, and iterate to achieve production‑grade reliability.
Requirements
- Strong programming skills in Python and C++ with experience in embedded development.
- Hands‑on experience building and optimizing audio ML models (e.g., speech enhancement, acoustic echo cancellation, sound event detection).
- Proficiency in model optimization techniques such as quantization, pruning, and hardware‑specific acceleration.
- Familiarity with Snapdragon NPU architecture and toolchains for on‑device deployment.
- Solid understanding of signal processing fundamentals and real‑time system constraints.
Skills
pythonctensorflowpytorch