Interview: September 28, 2026 · US Pacific Time · Approximately 54 minutes
Interviewed and edited by Damnang · Damnang Research
Expert: Meta advertising software engineer
When I look at Meta, I usually focus on how much computing capacity it needs to develop AI models and what that investment means for semiconductor demand. I had spent less time looking at what AI already does in Meta’s main business, advertising. I wanted to understand what better models change for advertisers and whether those improvements actually lead them to spend more.
Using Muse recently gave me more questions. Beyond handling email and organizing my schedule, it could open websites and do things I would otherwise have to do myself. That was what caught my attention. If people use services like this regularly, how might Meta’s existing business benefit? I wanted to explore the connection to advertising and the experiences that might become possible alongside products such as smart glasses.
The expert I spoke with is a software engineer at Meta who supports large advertisers. The role involves integrating capabilities developed by the AI team into advertising products and adapting them to advertisers’ objectives, rather than researching the models that predict how people will respond to ads.
Later in the conversation, we moved from AI in software development to Muse and smart glasses. Could a service I find useful in daily life also help Meta’s advertising business? What might change when smart glasses enter the picture? I wanted to hear the perspective of an engineer working in advertising.
Contents
01 Connecting Advertising AI to Advertiser Goals
02 How an Ad Slot Is Filled
03 How Meta Understands User Behavior
04 Small Performance Gains Can Change Advertiser Decisions
05 Buying After Seeing an Ad Versus Buying Because of It
06 Why Getting Started Is Harder for Small Advertisers
07 People Still Need to Set Business Goals
08 Extending Support Beyond Large Advertisers
09 How an Advertising Engineer Uses AI
10 Using Muse and Using Its Data for Advertising
11 Why the Expert Is More Interested in Smart Glasses
12 Analysis by Damnang
Editorial note: This Q&A combines written responses provided in advance with a subsequent interview. Related remarks have been reordered, and wording and repetition have been edited. It is not a verbatim transcript. The expert’s views are distinct from Meta’s official position. Explanations added from public sources are labeled Editor’s note, while my own assessment appears in Analysis by Damnang.
01 Connecting Advertising AI to Advertiser Goals
Damnang: Could you start by introducing yourself? You work in advertising at Meta. What does your role involve?
Expert: I’m a software engineer working on advertising at Meta. Our team supports large advertisers. Even when companies use the same Meta advertising platform, what they want can be quite different. Some want to focus on photo ads, while others prefer video. They sell different products to different customers, so their requests can get quite complicated.
They also want different things from their advertising. Some campaigns aim to generate purchases right away. Others aim to get people to install an app or add a product to their cart. We develop and support the features that let advertisers use Meta’s capabilities for those goals.
Damnang: Does that mean you also build the algorithms that decide which ads to show to which users?
Expert: That part is a little different. A separate AI team builds the technology that predicts what a user will do after seeing an ad. My work is more about connecting the capabilities that team develops to the actual advertising products.
Damnang: Could you give an example of what that involves?
Expert: Let’s say a clothing retailer advertises a collection of fall products. There might be sweaters, pants, and various other items. There’s a feature that optimizes which product to show to each user. For this example, imagine it was originally available only for ads with a purchase objective.
The advertiser might want to use that feature in other campaigns too. This time, they want more people to add items to a cart or a wishlist, rather than make a purchase. Someone then needs to make that same optimization feature work for those objectives. That’s closer to the kind of work I do.