Workday’s CTO traded his C-suite title for a technical staff role at Anthropic

In short: Peter Bailis, who joined Workday as chief technology officer in May 2025, left the company last month and has taken a role as member of technical staff at Anthropic, where he will focus on reinforcement learning engineering. The move strips away a C-suite title in exchange for technical proximity to the frontier, and lands Bailis inside a company that is now openly building the kind of HR software that Workday sells.

The hire was confirmed by an Anthropic spokesperson and reported by Business Insider. Bailis’s profile on Workday’s leadership pages has been removed, and his LinkedIn profile reflects the transition. Before joining Workday, Bailis served as vice president of engineering at Google Cloud, where he led AI for data initiatives including Conversational Analytics, NL2SQL, and retrieval-augmented generation for structured data. Prior to that he founded and ran Sisu Data, a decision intelligence company, and was an assistant professor of computer science at Stanford University, where he co-led the DAWN project on data-intensive applications. His research background in data systems and his recent enterprise AI work at Google Cloud make him an unusual profile for a reinforcement learning hire — and one that signals Anthropic’s interest in engineering that bridges model training and applied product development.

From C-suite to individual contributor

The title Bailis has accepted at Anthropic, member of technical staff, or MTS, is the standard engineering designation at both Anthropic and OpenAI, applied across research and engineering regardless of seniority. OpenAI president Greg Brockman has explained the structure publicly, saying the labs did not want to “bucket people into researchers and engineers,” treating the MTS title as a signal of flat technical hierarchy rather than an organisational step down. In practice, the title spans a very wide range of seniority and compensation: MTS base salaries at Anthropic run from approximately $300,000 to $405,000, while at OpenAI the band stretches from roughly $210,000 to $530,000, with equity grants pushing total compensation well into seven figures for senior positions.

For Bailis, the trade-off is explicit. A CTO role at a company of Workday’s scale, roughly $8 billion in annual revenue, 18,000 employees — carries institutional authority of a kind that an individual contributor role at even a frontier AI lab does not. What it does carry, however, is directness: access to the models being built, the training decisions being made, and the research agenda being set, without the layers of organisational management that sit between a CTO and the actual work. Anthropic, which now reports a revenue run rate exceeding $30 billion and has more than 1,000 enterprise customers each spending over $1 million annually, is no longer a research lab that happens to generate revenue — and for a technically ambitious executive, the scale of the engineering questions it is now confronting is a draw in itself.

What makes Bailis useful to Anthropic right now

The Anthropic spokesperson confirmed that Bailis will work on reinforcement learning engineering. That is a broad remit at a company where reinforcement learning from human feedback underpins model alignment, and where RL-based training techniques are also being applied to agent behaviour and tool use. What Bailis brings beyond general engineering capability is deep familiarity with the enterprise software stack that Anthropic is now moving into as a product company, not merely as an API provider.

Specifically, The Information reported that Bailis’s arrival coincides with Anthropic’s push to build HR applications. Anthropic has already launched Claude plugins for HR use cases, including generating job descriptions, onboarding materials, and offer letters. Bailis spent the past several years inside the enterprise HR and finance software world, first at Google Cloud, where he built AI products that plugged into structured enterprise data, and then as Workday’s CTO, where he led the company’s agentic AI strategy. That domain expertise is now being pointed at Anthropic’s enterprise product roadmap from the inside.

The competitive irony is pointed. Workday’s CEO, Carl Eschenbach, stated publicly in February 2026 that Anthropic, Google, and OpenAI all use Workday’s software internally. The company that bought licences from Workday has now hired its CTO to build products that will compete for the same enterprise HR budgets.

Anthropic’s enterprise push

Bailis’s hire makes more sense in the context of how aggressively Anthropic has been building enterprise distribution infrastructure over the past several months. In early March, Anthropic launched a marketplace for Claude-powered enterprise software, allowing companies with committed API spend to purchase third-party applications built on Claude, with Anthropic foregoing the revenue cut that cloud hyperscalers typically charge. Launch partners included Snowflake, legal AI firm Harvey, and developer platform Replit.

Days later, Anthropic committed $100 million to a new Claude Partner Network with Accenture, Deloitte, Cognizant, and Infosys as anchors, formalising consulting relationships designed to accelerate Claude deployments inside the world’s largest enterprises. The company said it expected to increase that commitment in subsequent years and was scaling its partner-facing headcount fivefold. Separately, Anthropic has been in talks to anchor a joint venture with private equity firms including Blackstone that would embed Claude across portfolio companies in exchange for roughly $200 million of Anthropic’s own capital.

Enterprise customers now represent approximately 80% of Anthropic’s revenue, and the company’s head of enterprise product has described its positioning as a platform rather than a product — one that aims to operate inside existing enterprise workflows rather than to displace the software those workflows run on. In practice, the distinction is difficult to maintain cleanly: if Claude can generate offer letters, onboarding packs, and job descriptions, it is beginning to substitute for the value that HR software platforms deliver, even if it sits technically alongside them.

The talent signal

Bailis is not the first senior enterprise technology executive to trade a corporate title for an individual contributor role at a frontier AI lab. The pattern has been described by observers as rare but growing: executives at the top of established technology companies choosing proximity to cutting-edge research over the authority that comes with managing large organisations. The motivation, where it has been articulated, tends to be the same — a sense that the most consequential technical work is happening at AI labs, and that watching it from a CTO seat at an enterprise software company is a worse position than doing it.

For companies like Workday, the pattern presents a specific challenge. The CTO role is a competitive signal as much as an operational one — it attracts technical talent, shapes engineering culture, and communicates to the market that the company is at the frontier of what it does. Established enterprise software companies restructuring their leadership and technical strategies in the face of AI disruption is now a recurring story. Workday has not announced a replacement for Bailis. The company is currently rolling out its own Agent Builder tools, which allow enterprise customers to build AI agents on top of Workday’s data, and the CTO position is central to positioning that product as differentiated from the Claude-based alternatives that Anthropic is now building with Bailis’s help.