Meta is setting up a new applied AI engineering organisation as part of its broader push toward building advanced “superintelligence” systems, according to a recent report.
The new group will be led by Maher Saba, who currently serves as a vice-president within Meta’s Reality Labs division. The team will report directly to Andrew Bosworth, the company’s Chief Technology Officer. Internally, the structure is expected to be unusually flat, with as many as 50 employees reporting to a single manager.
The unit is designed to work alongside Meta’s Superintelligence Lab, which was launched last year as part of a major AI strategy overhaul. In an internal memo, Saba described the new organisation as building the “data engine” that helps improve AI models more quickly and effectively.
The applied AI engineering group will be split into two main teams. One will focus on developing tools and interfaces. The other will handle task execution, data generation and evaluations — feeding real-world performance insights back into the model development process.
Saba noted that building strong AI models requires more than just research talent and computing power. Real-world data, continuous feedback and structured evaluation systems are essential to creating a feedback loop that steadily improves performance. He pointed to recent gains from reinforcement learning and post-training methods, suggesting that accelerating these efforts could give Meta an edge.
Meta’s renewed AI push began last summer with the formation of the Superintelligence Labs, led by Alexandr Wang, formerly the chief executive of Scale AI. That team is responsible for developing Meta’s next-generation AI models.
Since then, the company has been ramping up hiring and working on new models internally codenamed Avocado and Mango. Investors have been closely watching progress, with CEO Mark Zuckerberg facing questions about the company’s AI roadmap during earnings calls.
In January, Zuckerberg said Meta would begin rolling out its new models and AI-driven products in the coming months. He indicated that while the first releases may not be perfect, they would demonstrate the pace at which the company is advancing.
The newly formed applied AI engineering team is now recruiting managers from across software engineering, product management, design, data science and data engineering. Interested employees have been asked to submit an internal application form by the end of the week.
Also Read: Meta Finalises Multi-Billion Dollar Agreement With Google for AI Chips
