Applied Scientist, Silicon and Systems Group Edge AI
Core
Develop new evaluation methods and automated reliability validation for multimodal language models and agents on Amazon devices (audio, vision, video).
Role type
Applied Scientist (Edge AI, Multimodal Model Evaluation)
Builds
Evaluation frameworks and automated testing methods for multimodal AI models on consumer hardware.
Domain
Consumer Electronics / Edge AI / Multimodal Machine Learning
Deliverable
production ML models
Required skills
Multimodal language models, automated evaluation methods, LLM-as-judge, perception tasks, Java, C++, Python, algorithms and data structures, numerical optimization, parallel and distributed computing, high-performance computing
Preferred skills
Unix/Linux, professional software development
Technologies
Java, C++, Python, LLM-as-judge
Responsibilities
Invent and validate reliability for novel automated evaluation methods for perception tasks; Develop and extend evaluation frameworks for multimodal language models; Analyze large offline and online datasets to interpret model failures; Collaborate with training teams to enhance model capabilities for product use cases.
Seniority
Mid-Senior, hands-on IC