Principal Applied Scientist

Responsibilities

• Design and implement novel deep learning architectures combining multiple modalities, including image, video, and geospatial data.
• Solve computational problems to train foundation models on vast datasets and enable large-scale inference.
• Collaborate with science and engineering teams to build adaptations for Amazon Last Mile deliveries, improving safety and efficiency.
• Guide technical direction for research initiatives and ensure robust performance in production environments.
• Mentor fellow scientists and engineers while maintaining significant hands-on technical contributions.
• Lead technical initiatives from conception through deployment and drive discussions with key stakeholders.

Requirements

• PhD or Master's degree and 8+ years of applied machine learning experience.
• Experience designing novel deep learning model architectures and building models from scratch.
• Proficiency with data tools, including SQL and Spark.
• Expert coding skills in Python, C++, or other production-level languages.
• Strong publication record at top-tier conferences (e.g., NeurIPS, ICML, ICLR, CVPR) or demonstrated industry innovation in ML.
• Extensive track record of leading technical projects and mentoring junior staff.

Preferred Qualifications

• Experience building foundation models for industry or research applications.
• Experience designing multi-modal model architectures.
• Experience building models for motion prediction, such as for autonomous driving.
• Proven track record of successful production ML deployments.
• Experience with large-scale distributed environments for ML training and inference.

Benefits

• Competitive base salary and sign-on payments.
• Restricted Stock Units (RSUs).
• Comprehensive health insurance (medical, dental, vision, and prescription).
• 401(k) matching and paid time off.
• Parental leave and mental health support.

About the Company

Amazon's Delivery Foundation Model team combines ambitious research vision with real-world impact. We leverage unparalleled computational infrastructure and extensive datasets to deploy state-of-the-art foundation models that improve the safety, quality, and efficiency of Amazon's global Last Mile delivery network.