Get matched free

Applied Scientist, Selling Partner Support Engagement

Amazon (AI roles) · Vancouver, British Columbia, CAN

No salary listed. Estimated from 401 salary-disclosed Research roles on this board: $225K–$330K (interquartile range; estimate, not the employer's figure).

Apply at Amazon (AI roles)Find your warm intro on LinkedIn
About the role
Amazon's Selling Partner Support handles tens of millions of contacts annually worldwide. Our science team is transforming this experience by building AI agents that autonomously resolve seller issues, learn from every interaction, and continuously improve with minimal human intervention. These agents reason and adapt — from understanding the seller's context and selecting the right solution, to routing contacts optimally and automating resolution end-to-end. We do this in deep partnership with multiple engineering and product teams.

We are looking for an Applied Scientist who is excited to work on reinforcement learning, agentic architectures, and large-scale production systems. You will be directly connected to the problems sellers face every day, translating real customer pain into science solutions that operate at massive scale. You will frame ambiguous business challenges as tractable machine learning problems and ship systems that measurably improve millions of seller interactions.

Key job responsibilities
- Optimize cost of agents in runtime by exploring deployment and domain adaptation of System One models.
- Own research and development of reinforcement-learning based agent improvement systems, from problem formulation through production deployment and impact measurement.
- Design approaches to preference learning, reward modeling, and policy optimization for conversational agents operating over real-world tools and APIs.
- Build evaluation frameworks that measure agent quality across dimensions such as helpfulness, correctness, and safety.
- Collaborate with engineering and product teams to deploy models into production systems serving sellers worldwide.
- Contribute to the team's scientific impact through peer review, mentorship, and publications at peer-reviewed venues.

A day in the life
You read the latest research papers and implement new techniques by building rapid prototypes using AI-assisted coding tools, then take what works from prototype to production. You collaborate closely with product managers and engineering teams to translate seller pain points into deployed science solutions. You work with senior scientists to propose experiments and make the case for what to build next. You also stay connected to the seller experience by understanding the real problems sellers face so your models solve what actually matters.

About the team
Our team is a growing group of scientists building the AI that powers Amazon's seller support experience. We operate across capabilities such as agentic systems, knowledge retrieval, and content automation, each owning distinct problem spaces but sharing evaluation infrastructure and research insights. We work backwards from business problems, deeply understanding the domain and defining success metrics and guardrails. This lets us run parallel experiments, compare approaches rigorously, and ship the best models to production. We publish at both internal and external venues and actively invest in research that compounds over multiple product cycles.

The team sits in Seattle/Vancouver and operates with high autonomy. Scientists own their domains end-to-end, from problem framing through production deployment. We value speed over perfection and experimentation over debate. We value inclusive experiences. Even if you do not meet all of the preferred qualifications listed above, we encourage you to apply. The team supports an inclusive learning culture where individual growth is a priority — you will find mentorship, knowledge-sharing, and career-advancing resources here. - Experience programming in Java, C++, Python or related language
- Experience with SQL and an RDBMS (e.g., Oracle) or Data Warehouse
- PhD in Engineering, Computer Science, Machine Learning, Operations Research, Statistics, or related fields
More like this

More Research roles · All Amazon (AI roles) jobs

Is this your role?
Claim it free with your work email to see how it’s performing here: views, apply-clicks and saves from candidates browsing AI roles. You can also keep it up to date or mark it filled.
We create your account on first use. Only an email on the company’s own domain can claim a role.