AirDNA Innovation & Technology Culture

How does innovation show up in your company culture?

I deeply believe innovation can come from anywhere. Very rarely do the best ideas come from the key decision-makers. They usually come from places like the customer experience team, who are on the frontlines talking to customers every day, or from the quietest person in the room.

Innovation is also hard. It’s hard because you have to force yourself to think in terms of first principles. You have to zoom out and ask questions that feel obvious in hindsight, but no one slows down to ask. We spend a lot of time in the weeds, but it’s important to spend time in the clouds, too. At AirDNA, we are intentional about giving people the time and permission to question assumptions and rethink how we do things.

What’s one recent innovation that improved user or employee experience?

We are leaning into AI like most companies right now. A big part of our business is helping people project how much an Airbnb could make before they buy it. We do this by identifying nearby properties that are already operating as short-term rentals. The more precise we are in selecting these “comps,” the more accurate the revenue projection.

Our data team has spent years refining this process and making AirDNA best-in-class. AI gives us new ways to push it further. Right now, it’s internal. Soon, it’s going to fundamentally change how investors predict revenue.

How do you balance experimentation with stability?

This is kind of a trick question. Effective experimentation is not “stable” by some definition. If you want meaningful wins, you have to try things that are … experimental. If you optimize for playing it safe, there’s a good chance you never learn anything. It’s easy to say but hard to practice.

When it comes to technical stability, there are excellent tools that let you experiment with guardrails in place. At AirDNA, we use Statsig to run controlled experiments and monitor impact so we can move fast without being reckless. You still have to be willing to take real swings though. That is where the upside lives.

What’s it like to work on the AI and machine learning team at your company?

Working on the AI and ML team here is a fun challenge and a lot of that comes down to the nature of our data. We’re not pulling clean rows out of a database. We’re scraping the short-term rental market across Airbnb, VRBO and Booking.com and each platform has its own quirks, gaps and ways of changing overnight. A listing’s status can flip, a property can disappear, prices can spike for reasons that have nothing to do with the market. 

That means a big part of the work isn’t just modeling. It’s figuring out what the data is telling us and then deciding what to trust. Almost every problem starts as a question about the data before it becomes a question about the model.

The modeling side is where the variety kicks in. On any given week we might train a classification model to infer a listing’s status, work through pretty much every flavor of regression to forecast demand or pricing and turn around and use state-of-the-art LLMs on a completely different project. No single tool fits. The problem dictates the approach and that’s what keeps it fun.

How is your team applying emerging technology in practical, business-relevant ways?

Our approach to emerging tech is pretty simple. We don’t jump on the hype train just because something is new. There’s no shortage of cool tools out there, but cool doesn’t always mean useful and useful doesn’t always mean useful for us.

Before we pick up a new technology, we take the time to understand it. What problem does it solve? Where would it fit in our stack? Is it solving something we already have, or are we inventing a problem to justify the tool? Those questions sound obvious, but skipping them is how teams end up with shiny things that don’t move the needle.

When something does pass that bar, we lean in. We’ve integrated LLMs into projects where they genuinely outperform what we had before and we’ve passed on plenty of other ideas where the math didn’t work out. The goal is always the same; figure out where emerging tech can give us a real edge on the data and problems that are unique to short-term rentals. That’s where differentiation comes from, not from being early for the sake of being early.

What should candidates know about the tools, collaboration or problem-solving involved in AI work at your company?

The biggest thing for candidates to know is that every idea is welcome here. Whether it comes from someone with ten years of ML experience or someone who just joined, we’d rather hear the idea and figure out together if it has legs than miss it because the room wasn’t open to it.

A lot of our current AI work is focused on automating processes that have historically been very manual. When your data is messy and your workflows have grown organically over time, there’s plenty of room to make people’s days easier and free them up for higher-leverage work. That’s where a lot of our energy is going.

What ties it all together is collaboration. Engineering, product and data each see a different piece of the puzzle and the best ideas usually come out of those conversations rather than from any one person sitting alone with a model. The candidates who do well here are the ones who like that back-and-forth and who care as much about whether something is useful as whether it’s interesting.