Overview As a Senior Applied Scientist in Titans Science, you will translate real seller pain into scalable AI solutions at massive scale. You work at the intersection of reinforcement learning, agentic architectures, and production systems to improve how seller issues are resolved. You will own end-to-end ML research and deploy models that impact millions of interactions, collaborating across multiple engineering teams and influencing strategic investments. This role combines rigorous science with practical impact in a fast-moving, autonomous team at Amazon Seattle.
Compensation / Benefits- health insurance (medical, dental, vision, prescription)
- 401(k) matching
- paid time off
- parental leave
- RSUs
- sign-on payments
Responsibilities- Own end-to-end RL-based agent improvement research from problem framing to deployment and impact measurement
- Design approaches for preference learning, reward modeling, and policy optimization for conversational agents
- Build and maintain evaluation frameworks assessing agent quality across helpfulness, correctness, safety, and alignment
- Collaborate with scientists in NLP, optimization, ML, and statistics; partner with 10+ engineering teams to deploy models
- Publish research at NeurIPS, ICML, EMNLP, AMLC and contribute to hiring and mentoring
Key requirements- 7+ years of applied research experience
- 5+ years of ML model-building for business applications
- PhD, or Master's degree with 5+ years of applied research experience
- Programming experience in Java, C++, Python or related languages
- Experience with neural deep learning methods and ML
- collaboration with cross-functional teams
- mentorship of junior scientists
- rigor and execution in research
- RL and agent-based systems
- NLP and knowledge retrieval
- policy optimization and reward modeling
