Applied Scientist — Ads Forecasting & Insights (FIG)

Amazon Inc.; Palo Alto (CA); On-site; USD 172,000 - 222,000; Full time

Job summary

Amazon.com Services LLC in Palo Alto is seeking an Applied Scientist to advance advertising through ML and generative AI. You will take models from problem formulation to production, collaborating with engineering and product teams to deliver measurable advertiser impact.

Ideal candidates have hands-on experience with large language models, sequential decision-making, causal inference, and modern DL frameworks like PyTorch. You will lead experiments and ship science into customer-facing products.

Qualifications

• 3+ years of building models for business applications.
• PhD or MSc with 4+ years in CS/CE/ML.
• Proficient in Java, C++, or Python.
• Experience with algorithms, data mining, or HPC.
• Experience training or adapting large language models.
• Experience with sequential decision-making or optimization under uncertainty.
• Experience with causal inference or rigorous evaluation of ML systems.
• Experience with PyTorch.

Responsibilities

• Lead end-to-end ML projects from framing to production deployment.
• Develop models to improve advertiser outcomes on Amazon ads.
• Design experiments and statistical analyses to validate impact.
• Collaborate with engineering and product to ship AI into products.
• Advance the state of the art by researching new techniques for advertising.

Skills

Model building

Programming languages: Python/Java/C++

Algorithms & data structures

Numerical optimization

Data mining

Parallel & distributed computing

High-performance computing

ML research / LLMs

Experiment design

Causal inference

Education

PhD

MS with 4+ years CS/CE/ML

Tools

PyTorch

C++

Java

Python

Job description

Amazon.com Services LLC in Palo Alto is seeking an Applied Scientist to advance advertising through ML and generative AI. You will take models from problem formulation to production, collaborating with engineering and product teams to deliver measurable advertiser impact.

Ideal candidates have hands-on experience with large language models, sequential decision-making, causal inference, and modern DL frameworks like PyTorch. You will lead experiments and ship science into customer-facing products.