Applied Scientist, Amazon Ads

Saved job description

Amazon Ads runs one of the largest real-time decision systems in the world. To meet advertiser goals, we process >4 million ad requests per second, and this team decides which of them are worth the compute. We mine a petabyte-scale stream of traffic to tell in-demand, valuable requests apart from wasteful ones, filtering out the noise and forwarding only what brings value to advertisers, all within milli-seconds. Our solutions save millions in compute costs and drive long-term growth for Amazon's Retail and Marketplace businesses.

Why you'll love this role: The scientific challenge here is scale and signal: large-scale ML on enormous, high-velocity data, optimization and auction dynamics, and causal measurement of what actually moves the business. Generative AI and LLMs are tools you reach for when they help, not the point of the work. The impact also reaches beyond Amazon: we donated the Dynamic Traffic Engine (DTE) to the IAB Tech Lab as an open-source framework that lets demand-side platforms signal which bid requests they value, so supply-side platforms send fewer wasteful ones and QPS waste drops across the whole programmatic supply chain. We are a highly motivated, collaborative team with an entrepreneurial spirit and a broad mandate to experiment and innovate. You will work alongside talented engineers, product managers, and scientists in a culture that rewards bias for action, innovation, and experimentation. What makes this role unique is the combination of scientific rigor with real-world impact at a scale very few systems reach.

Your impact and career growth: Amazon Ads is investing heavily in AI and ML, creating room for scientists to innovate and make their mark. As an Applied Scientist, you will own the science for well-defined supply-optimization problems and take your work end to end, from framing the approach through experimentation and into production with measured business impact. You will partner closely with engineers to productionize your models and with product managers to tie your science to customer and advertiser outcomes. You will share your findings, begin to mentor teammates, and grow your influence on the team's technical direction over time. This is your opportunity to do high-rigor applied science in one of the fastest-growing businesses at Amazon.

Learn more about Amazon Ads: http[local path removed]

Key job responsibilities

As an Applied Scientist in Amazon Ads, you will:

- Conduct hands-on data analysis and build large-scale ML models and pipelines for supply optimization.

- Take your models end to end, from proof of concept through experimentation, optimization, and deployment into production, working closely with software engineers to productionize them and with product managers and fellow scientists along the way.

- Design and run A/B experiments and causal analysis to validate hypotheses, draw insights from petabyte-scale data, and measure real business impact.

- Research and apply advanced ML, including generative AI and large language models, to demand prediction and traffic filtering.

- Drive end-to-end optimization projects that tackle ambiguous problems at massive scale, developing scalable, efficient processes for model development, validation, and deployment that optimize traffic monetization while maintaining customer experience.

Requirements: - 3+ years of building models for business application experience

- PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience

- Experience programming in Java, C++, Python or related language

- Experience in any of the following areas: algorithms and data structures, parsing, numerical optimization, data mining, parallel and distributed computing, high-performance computing