Mar 4, 1:08 PM
Meta Launches New Applied AI Engineering Unit to Accelerate Superintelligence Push
New teams led by Maher Saba will operate with a flat structure to speed development of next-generation AI models
Meta Platforms is setting up a new Applied AI Engineering team aimed at ramping up its push to create cutting-edge artificial intelligence systems.
Maher Saba will head this fresh unit, which will answer directly to Meta’s chief technology officer, Andrew Bosworth. The team setup will be unusually flat—imagine up to 50 engineers or specialists reporting to a single manager. This group will work hand-in-hand with Meta’s Superintelligence Labs, the outfit behind the company’s most sophisticated AI models.
Saba’s memo outlines the mission: to build the “data engine” that speeds up improvements in AI by generating feedback, evaluations, and real-world data.
The new squad divides into two main teams. One crafts tools and interfaces for AI development; the other zeroes in on task execution, generating training data, and model performance assessment. That info loops back to Meta’s modeling groups to refine the systems further.
“Building great models isn’t just about researchers and compute,” Saba noted. “It requires real-world data, feedback and evaluations that help turn a strong model into a leading one.” Meta shook up its AI efforts last year by founding the Superintelligence Labs division, led by Alexandr Wang, formerly the CEO of Scale AI. This team drives the development of Meta’s next-gen AI models.
Since then, the company has stepped up hiring and started work on several in-house models, internally code-named “Avocado” and “Mango.”
Investors and analysts are watching Meta's AI journey closely, especially amid fierce competition across the tech world.
On a recent earnings call, Meta CEO Mark Zuckerberg said the company plans to roll out the next wave of AI models and products soon.
“I expect our first models will be good,” Zuckerberg shared, “but the bigger play is showing how fast we can move.”
Those interested in leadership roles within the new Applied AI Engineering group have a deadline: apply internally by week’s end.
