Meet Amjad, Director of Applied Science for Sponsored Products and Brands

When Amjad joined Amazon Ads in 2019 as a Principal Applied Scientist, he had a number of offers from notable tech firms. His choice to join Amazon Ads was driven by the clear opportunity he saw to make an immediate impact. Since then, he's grown from the role of individual contributor into the role of a leader—first growing a team of engineers and scientists and then working as a Director in the Sponsored Products and Brands division.

I joined Amazon Ads in 2019 as a Principal Applied Scientist. I saw a clear opportunity to contribute from day one and make a significant impact, which was what interested me most. Given my background and where Ads was at that point, I understood how my expertise would immediately add value. I joined to focus on improving the relevance of Sponsored Products shown in Amazon search results, which was a hot area for Amazon Ads at the time. Over my five-year tenure with Amazon, I’ve contributed to multiple other areas, such as semantic retrieval, affinity modeling, search experience measurement, and most recently conversational shopping.

We practice what we call "customer-centric science." Amazon's Working Backwards approach means we start with specific customer problems and work to solve them rather than conduct pure research that might not lead to customer impact. Our teams are typically cross-functional, combining product, engineering, and science expertise, which enables better collaborations and faster implementation of solutions.

To start, there's an incredibly diverse range of problems that require different scientific expertise. Scientists work across information retrieval, natural language processing, image understanding, generative AI, game theory, auction mechanisms, and optimization. With the AI revolution, we're seeing unprecedented opportunities to make ads truly useful through better contextualization and explanation. Think about it—before, advertising was like a billboard with limited text, but now with AI, we can really explain why a sponsored recommendation is worth a customer’s consideration.

What's also unique is how we approach these challenges in a production environment. Building AI applications is one thing, but making them work at scale with real business constraints on cost and latency creates fascinating scientific problems. All our scientists have access to extensive AWS resources and tools, so they aren't limited by technical constraints; this is extremely helpful when working on challenging problems.

It’s fast-paced, and it can be challenging—but there’s a lot of support, and scientists can build a strong network of peers. Scientists are spread throughout Amazon, so we have a centralized Amazon Science program with the mission of supporting the entire science community. It provides opportunities such as mentoring, workshops, talks by top speakers from Amazon and the wider scientific community, and support for publications.

Amazon Science also organizes internal conferences, such as the Amazon Machine Learning Conference (AMLC). The AMLC is highly competitive, with an acceptance rate around 20%. It allows scientists across Amazon to stay up to date with state-of-the-art in various AI fields, to network with colleagues outside their immediate teams, and to initiate collaborations. It’s great to see scientists working on research projects across Amazon divisions.

Career development happens at multiple levels. There's a formal mentorship program where new scientists can connect with experienced mentors across different teams. Scientists can share their work at different stages through internal forums and get feedback from conception to deployment. And I think it’s particularly valuable that they have direct access to leadership who can engage in detailed technical discussions about their work.

Beyond Amazon, we actively support scientists in building their external profiles through conference attendance and participation in program committees. We also have dedicated teams focused on representing Amazon at top conferences, creating opportunities for scientists to engage with the broader scientific community.

Our writing culture aligns well with scientific thinking. Everyone from engineers to product managers write papers that require narratives supported by data and examples. We also have significant tolerance for iteration and experimentation. We understand that challenging problems require multiple attempts to solve; learning from unsuccessful experiments is valued as much as learning from successful ones.

We're not just thinking about advertising as a way to generate revenue; we put the customer first and focus on creating value, which is similar to Amazon’s overall approach. We view advertising as a discovery tool for both shoppers and advertisers. Scientists work across all layers of the challenge; this includes determining what ads to show and when as well as collaborating with UX designers on presentation. It's about making advertising truly helpful rather than intrusive.

We operate in more than 20 countries across different regions and cultures. Scientists can see their work launch in multiple markets, from Brazil to Singapore, impacting hundreds of millions of people worldwide. This global scale provides unique opportunities to solve complex problems across diverse contexts. The scale of the work we do is remarkable and very exciting.