🧠 Food for thought

Implications, context, and why it matters.

Meta is building an applied AI engineering “data engine” for its AI models

• Meta has formed a unit to run a continuous feedback loop that improves its AI models and supplies data for the company 1.
• The unit has two teams. One builds interfaces plus tools. The other completes tasks, creates data, and runs evaluations that go back to Meta’s AI model groups 1.
• The group sits in Reality Labs, Meta’s division for advanced technology projects. It reports to CTO Andrew Bosworth. The setup links applied AI engineering to Meta’s model training, evaluation, and improvement systems 1.

The future of engineering management could be extremely flat

• The restructuring uses a flat structure. Some engineering teams may reach manager-to-employee ratios near 1-to-50, according to employees familiar with the plan 1.
• This matches CEO Mark Zuckerberg’s push to elevate individual contributors and flatten engineering teams across Meta 1.
• The plan assumes AI can let one highly skilled software engineer finish work that used to need large teams. That shift would reshape engineering managers’ work. It would also change what individual software developers do 1.