Prominent artificial intelligence researcher and systems engineer Andrew Tulloch has departed Meta Platforms to join foundation model developer Anthropic, with industry reports citing a multi-year compensation package valued at approximately $1.5 billion.
The high-profile transition represents one of the largest recorded compensation commitments for an individual technical researcher in the technology sector, underscoring the escalating competition among frontier artificial intelligence laboratories to recruit elite engineering talent capable of training next-generation foundation models.
Tulloch, who spent several years working within Meta’s fundamental artificial intelligence research teams, has a recognized track record in distributed computing architectures, large-scale systems optimization, and parallelized training infrastructure. In the development of modern frontier models, scaling high-performance compute clusters across tens of thousands of specialized accelerators without communication bottlenecks or hardware failure interruptions is considered as critical as model algorithmic design.
Industry analysts observe that compensation packages of this magnitude are typically structured over multiple years, composed primarily of restricted equity units, vesting milestones, and performance incentives linked to corporate valuation growth rather than upfront cash payments. Anthropic, which develops the Claude series of multimodal models, has raised substantial capital from major corporate partners, including Amazon and Google, enabling it to match and exceed compensation levels traditionally offered by established Silicon Valley hyperscalers.
The transition highlights how systems-level optimization has become as vital as model architecture in frontier artificial intelligence development. As foundation models expand in parameter scale and training cluster requirements, researchers capable of maximizing accelerator utilization, minimizing interconnect latency, and preventing distributed hardware failures play a central role in controlling the immense capital costs associated with frontier training runs.