Applied Scientist, Live Data & Causal AI for Video

Amazon; San Francisco, Northern (CA, KY); Hybrid; USD 143,000 - 193,000; Full time

Job summary

Amazon Live Data Engineering, Sciences and Analytics (DESA) is seeking an Applied Scientist to build production-grade models, experiments, and signals that drive Amazon Live and Shop the Show metrics. You will own the full lifecycle from framing the research question through deployment and A/B experimentation, partnering with data engineers and a business intelligence owner to deliver measurable impact.

You'll design causal attribution and content intelligence models, develop ranking features,

Qualifications

• 3+ years of building models for business applications.
• PhD or Master's degree in CS/CE/ML with 4+ years in related fields.
• Experience with Java, C++, Python; PyTorch or JAX.
• Experience with ML fundamentals, training/inference lifecycles, and optimization.
• Experience with A/B testing, audience segmentation, and targeting.
• Experience with large-scale distributed systems (Hadoop, Spark).
• Experience with causal inference methods (incrementality, diff-in-diff, IV, synthetic control).

Responsibilities

• Design and deploy causal attribution models (incrementality testing, multi-touch).
• Build brand lifecycle models (LTV, CAC) and campaign optimization models.
• Design and run A/B experiments with proper methodology for new surfaces and ranking changes.
• Build multimodal and generative models for content intelligence at scale.
• Develop ranking and personalization features for downstream systems.
• Build predictive models proving video value to Amazon's systems.
• Own end-to-end lifecycle from research to production, monitoring, and iteration.
• Present findings to senior leadership and partner teams.
• Contribute to internal publications and cross-team methodology sharing.

Skills

Java

C++

Python

PyTorch

JAX

A/B testing

causal inference

distributed systems

Education

PhD or Master's in CS/CE/ML

Tools

Hadoop

Sparks

Bedrock

SageMaker

Redshift

Cradle

Job description

Amazon Live Data Engineering, Sciences and Analytics (DESA) is seeking an Applied Scientist to build production-grade models, experiments, and signals that drive Amazon Live and Shop the Show metrics. You will own the full lifecycle from framing the research question through deployment and A/B experimentation, partnering with data engineers and a business intelligence owner to deliver measurable impact.

You'll design causal attribution and content intelligence models, develop ranking features,