Senior Applied Scientist, APEX
Core
Building Alexa+, Amazon's LLM-powered conversational assistant, focusing on fine-tuning, alignment, agentic reasoning, and evaluation for hundreds of millions of customers.
Role type
Senior Applied Scientist (LLM & Agentic Systems)
Builds
Alexa+ conversational assistant, agentic AI systems, and evaluation frameworks
Domain
Consumer AI, Large Language Models, Conversational AI
Deliverable
production ML models
Required skills
LLM fine-tuning, RLHF/DPO alignment, agentic system design, distributed training, experimental design, Python, deep learning frameworks
Preferred skills
Agentic AI orchestration, large-scale model optimization, peer-reviewed publications, mentoring
Technologies
PyTorch, TensorFlow, Java, C++, R, scikit-learn, Spark MLLib, MxNet, numpy, scipy
Responsibilities
Design and implement novel approaches for LLM fine-tuning and alignment; Develop agentic systems with multi-step reasoning and tool use; Build evaluation frameworks for real-world conversational quality; Translate research into customer-facing products; Analyze large-scale experimental results; Publish results at top-tier venues; Mentor junior scientists
Seniority
Senior, hands-on IC with mentorship responsibilities