Applied Scientist, Alexa Connections
Core
Research, develop, and deploy next-generation on-device AI for communication agents that understand intent, context, and relationships to enable private, responsive interactions without relying on the cloud.
Role type
Applied Scientist (On-Device AI & Edge Intelligence)
Builds
On-device AI models, arbitration systems, and privacy-first personalization features for communication agents.
Domain
Consumer Technology / On-Device Machine Learning / Edge AI
Deliverable
production ML models
Required skills
Machine learning, deep learning, speech/NLU, on-device AI, model optimization (quantization, distillation, low-rank adaptation), C/C++, Python, Java, Perl, PyTorch, TensorFlow, on-device inference, LLMs
Preferred skills
PhD in CS/ML, on-device model optimization, on-device inference runtimes (ONNX Runtime, TensorFlow Lite, ExecuTorch), hardware accelerators (NPUs, DSPs, mobile GPUs), small language models (SLMs) for edge
Technologies
PyTorch, TensorFlow, ONNX Runtime, TensorFlow Lite, ExecuTorch, NPUs, DSPs, mobile GPUs
Responsibilities
Design on-device arbitration systems for routing between local and cloud inference, apply compression techniques to fit frontier-class capabilities into tight memory/latency budgets, optimize inference across on-device accelerators, build privacy-first personalization that learns without user data leaving the device, collaborate with hardware and systems teams to co-design models for target devices.
Seniority
Mid-level, hands-on IC
