Expedia’s AI Timeline Runs Through Its Talent Strategy

Expedia Group’s AI timeline increasingly depends on two related talent problems: finding engineers who get outsized gains from AI tools, and second, teaching the rest of the organization how they do it.

In an interview with Skift, Chief Technology Officer Ramana Thumu said Expedia spent nine months studying how its 5,000 engineers use frontier coding assistants and identifying top performers. “[It’s] not broad-stroke engineering productivity,” he said. “Top 2, 3, 5, 10% engineers, they’re exponentially more productive.”

At the same time, Thumu is recruiting AI, machine-learning, platform, cloud, and full-stack engineers to build the shared systems behind Expedia’s AI products. If Expedia fails to land enough of those engineers, projects Thumu wants completed in 12 months could stretch. Expedia has seen the clearest gains in retiring old systems and merging duplicate ones, the kind that piled up as it acquired companies over the years.