Top engineers choosing staff‑track roles
Peter Bailis left his Workday CTO role to join Anthropic as technical staff, highlighting a trend where senior engineers trade formal authority for deep technical leverage in AI infrastructure and RL work. The move signals that staff‑plus paths focused on connecting models to enterprise data can be a destination rather than a stepping stone to management (thenextweb.com).
A chief technology officer at a public software company just gave up the corner-office title to become “member of technical staff” at an artificial intelligence lab. Peter Bailis left Workday after joining as chief technology officer in May 2025, and Anthropic says he is now working on reinforcement learning engineering. (thenextweb.com) That title sounds smaller than chief technology officer, but at Anthropic it is the standard label for technical people across research and engineering. Anthropic’s careers pages list dozens of openings under artificial intelligence research, infrastructure, and software engineering rather than a ladder of flashy executive titles. (anthropic.com) The move makes more sense when you look at where the bottleneck is. In frontier artificial intelligence, the people closest to model training, evaluation, and deployment can shape the product more directly than someone managing a large org chart from above. (thenextweb.com) Bailis is not a generic executive crossing over late in his career. Before Workday, he led artificial intelligence for data work at Google Cloud, founded Sisu Data, and taught computer science at Stanford after building data-systems research in the DAWN project. (bailis.org) That background fits a very specific problem Anthropic is chasing. Large language models are good at generating text, but companies only trust them when they can pull the right facts from calendars, documents, databases, and internal software without making things up. (anthropic.com) Anthropic has been building the plumbing for that connection problem. In November 2024 it introduced the Model Context Protocol, an open standard for linking artificial intelligence assistants to business tools, content repositories, and development environments where company data already lives. (anthropic.com) Google then added official support for that protocol in its own services, which helped turn it from one company’s interface into shared infrastructure. When a standard starts spreading across vendors, the engineers who know both model behavior and enterprise data systems become unusually valuable. (cloud.google.com) Anthropic is also pushing harder into workplace software, not just chatbots. Its job board now spans 149 sales roles, 26 product engineering and design roles, 22 infrastructure software roles, and multiple openings tied to enterprise deployment. (anthropic.com) That is why a Workday veteran is interesting here. Workday sells the finance and human resources systems that hold employee records, approvals, payroll data, and planning workflows, and Anthropic is reportedly exploring human resources applications that sit much closer to that same data. (theinformation.com) Workday has already moved on by elevating Gabe Monroy to chief technology officer, but the more revealing part is what Bailis chose. In 2026, for some top engineers, the prestige job is no longer managing the people who build the engine; it is sitting next to the engine itself while the artificial intelligence stack is still being assembled. (erp.today)
