As a Research Analyst, you'll collaborate with experts to develop cutting-edge ML solutions for business needs. You'll drive product pilots, demonstrating innovative thinking and customer focus. You'll build scalable solutions, write high-quality code, and develop state-of-the-art ML models. You'll coordinate between science and software teams, optimizing solutions. The role requires thriving in ambiguous, fast-paced environments and working independently with ML models.

Key job responsibilities
• Collaborate with seasoned Applied Scientists and propose best in class ML solutions for business requirements
• Dive deep to drive product pilots, demonstrate think big and customer obsession LPs to steer the product roadmap
• Build scalable solutions in partnership with Applied Scientists by developing technical intuition to write high quality code and develop state of the art ML models utilizing most recent research breakthroughs in academia and industry
• Coordinate design efforts between Sciences and Software teams to deliver optimized solutions
• Ability to thrive in an ambiguous, uncertain and fast moving ML usecase developments.
• Familiar with ML models and work independent.
• Mentor Junior Research Analyst (RAs) and contribute to RA hiring


About the team
Retail Business Services Technology (RBS Tech) team develops the systems and science to accelerate Amazon’s flywheel. The team drives three core themes: 1) Find and Fix all customer and selling partner experience (CX and SPX) defects using technology, 2) Generate comprehensive insights for brand growth opportunities, and 3) Completely automate Stores tasks.
Our vision for MLOE is to achieve ML operational excellence across Amazon through continuous innovation, scalable infrastructure, and a data-driven approach to optimize value, efficiency, and reliability. We focus on key areas for enhancing machine learning operations: a) Model Evaluation: Expanding LLM-based audit platform to support multilingual and multimodal auditing. Developing an LLM-powered testing framework for conversational systems to automate the validation of conversational flows, ensuring scalable, accurate, and efficient end-to-end testing. b) Guardrails: Building common guardrail APIs that teams can integrate to detect and prevent egregious errors, knowledge grounding issues, PII breaches, and biases. c) Deployment Framework support LLM deployments and seamlessly integrate it with our release management processes.