According to a source with knowledge of the situation, Meta Platforms Inc. began executing the final phase of a three-part series of layoffs on Wednesday as part of a plan to eliminate 10,000 roles that was revealed in March.

After letting more than 11,000 employees go in the autumn, Meta became the first Big Tech business to announce a second round of mass layoffs in March. After a hiring frenzy that saw the company’s staff treble since 2020, the cuts reduced the company’s headcount to where it stood as of roughly mid-2021.

On Wednesday, several workers announced their layoffs on websites like LinkedIn, which was anticipated to have a significant impact on the ad sales, marketing, and partnerships departments.

Shares of Meta decreased 0.4% in premarket trading.

Mark Zuckerberg, the chief executive of Meta, stated in March that the majority of the company’s second round of layoffs would occur in three “moments” over the course of several months, mostly wrapping up in May. After that, he suggested continuing with a few smaller rounds.

Overall, the layoffs disproportionately affected non-engineering positions, highlighting the importance of Meta’s code writers. In March, Zuckerberg promised to “substantially” reorganize business teams and revert to a “more optimal ratio of engineers to other roles.”

According to executives who spoke at a subsequent business town hall, even amid layoffs intended particularly at technical teams, the corporation significantly slashed non-engineering functions like content design and user experience research.

Following a smaller blow to recruiting teams in March, some 4,000 people lost their jobs in the April layoffs, according to Zuckerberg during the town hall.

Layoffs at Meta came after months of declining revenue growth caused by rising inflation and a decline in digital advertising due to the widespread e-commerce expansion.

The business has also been investing enormous sums of money on Reality Labs, a division focused on the metaverse that lost $13.7 billion in 2022, as well as an effort to revamp its infrastructure to enable AI research.