Applied Scientist, Advertising

Applied Scientist, Advertising Overview

This role sits within Amazon’s demand retrieval function in the Amazon DSP organization, where the team tackles the challenge of selecting the most relevant advertisement from an enormous pool while respecting advertiser targeting goals and preserving a strong shopper experience. The work happens at very high scale and very low latency, and the team develops and runs machine learning systems that connect ad opportunities with the best matching ads so the right message reaches the right customer at the right moment.

The position is focused on improving ad matching for Amazon’s programmatic advertising products. In this role, you will help design solutions for performance sourcing by using behavioural signals from customer interactions with Amazon and other owned and operated properties, along with contextual details from bid requests, to estimate conversion propensity and improve campaign outcomes. The work has direct impact on very large businesses, and the team expects the scientist to contribute ideas, build significant technical advances, and help deliver meaningful improvements to Amazon’s advertising performance.

What you will do

• Develop and implement deep learning approaches that improve how customers are matched to ads across different business categories, countries or regions, and ad formats.
• Investigate and apply advanced machine learning methods, including multi-task learning, so the models can support advertisers from many industries and with different amounts of conversion data.
• Raise model quality by improving generalisation, scalability, and overall performance through better feature design and architectural enhancements.
• Collaborate directly with engineers to deliver production code changes in the advertising stack, working with very large datasets and systems that must handle high throughput.
• Rapidly test multiple hypotheses and implementation options in a setting where requirements may be uncertain, combining quantitative evidence with practical business judgement.
• Build a strong understanding of Amazon advertisers and their goals, then use that understanding to shape future products and technical capabilities over the long term.
• Keep up with the latest research in machine learning for recommender systems and advertising, and use that knowledge to inform the team’s strategic direction and investment choices.
• Work closely with product and engineering partners, contributing ideas, aligning on priorities, and adapting plans as business needs and experiment results evolve.
• Apply theoretical knowledge of modern machine learning methods together with software engineering and data processing skills to deploy solutions at Amazon’s advertising scale.

Team context

The demand retrieval team is responsible for building, deploying, and operating machine learning models that match bid opportunities to advertising demand based on performance signals, campaign delivery needs, and the targeting goals set by advertisers. The team evaluates success through offline experiments and online business metrics that reflect the effect of matching models on campaign performance indicators such as cost per action, return on ad spend, budget delivery, and targeting precision.

What the team is looking for

• A candidate with strong technical depth and the ability to work effectively with others.
• Someone who communicates clearly and works well across disciplines.
• A person who is motivated by business outcomes and comfortable operating in a fast-paced environment.
• Experience or familiarity with retrieval and ranking systems used in advertising or recommender systems is especially valuable.

Minimum qualifications

• A PhD, or a master’s degree combined with research experience in computer science, computer engineering, machine learning, or a related field.
• Programming experience in Java, C++, Python, or a comparable language.
• Experience creating machine learning models for business applications.
• Experience with modern deep learning model architecture design, deep learning training and optimisation, and model pruning.

Additional information

Amazon says it is an equal opportunities employer and makes hiring decisions based on experience and skills. The company states that it values diversity, inclusion, and a culture that helps employees deliver strong results for customers. It also notes that it does not discriminate on the basis of protected veteran status, disability, or other legally protected status.

If you need a workplace accommodation or adjustment during the application, interview, or onboarding process, Amazon directs you to its accommodations information page for support. The posting also refers applicants to Amazon’s privacy notice for details on how candidate data is collected, used, and transferred.

The listing is for the London location and is part of the Machine Learning Science job family.

Degree Requirement: Bachelor’s

Visa Sponsorship May be