AI Labs Job Tracker — Week of May 4–11, 2026

Thirty-eight confirmed new postings. Seven labs. One week. Here's everything that moved on the frontier AI hiring front between May 4 and May 11, 2026.

Three things stand out this week: Anthropic's 27-role surge is the highest single-week count across all monitored labs; xAI is spending more of its hiring budget on Memphis concrete and switchgear than on software; and OpenAI quietly opened an ads monetization cluster that signals ChatGPT is heading toward a revenue diversification push.

OpenAI's 655-total-open-position board 1 picked up 22 new listings over the past seven days. The postings fall into four clusters.

Safety and research. Two roles signal active investment in alignment infrastructure. A Researcher, Alignment Oversight 2 opening in San Francisco (hybrid, $250K–$445K + equity) focuses on control, accountability, and alignment for increasingly agentic AI systems. One day earlier, a Researcher, Recursive Self-Improvement Preparedness 3 role appeared ($295K–$445K), covering AI R&D risk measurement, data-poisoning mitigations, and model behavior science. Both sit in OpenAI's Preparedness / Alignment org.

Ads and revenue operations. Three postings in two days suggest an ads business build-out: an Ads Product Manager 4 in the ChatGPT Ads team (May 7), a Revenue Operations Lead, Ads Solutions 5 (May 6), and an AI Workflow & Operations, Revenue Strategy & Operations 6 role ($239K–$265K, May 7). Salaries for the Ads PM and Revenue Ops Lead were not available from the postings.

Commercial expansion. A Private Equity Lead 7 ($335K–$350K + equity, New York City) builds out OpenAI's PE commercial strategy. An APAC Sales Development Leader 8 in Singapore and a Lead, AI Deployment Manager 9 (also Singapore) both expand the Asia-Pacific footprint. A Partnerships Lead, Healthcare 10 rounds out the sector push.

Engineering and platform. Two Applied Foundations engineers — Full-Stack 11 and Backend 12 ($185K–$385K each) — will build identity, access, safety, payments, and commerce layers. A Software Engineer, Productivity - Inference Runtime 13 ($230K–$385K) supports inference infrastructure teams.

Other roles this week include an Executive Transformation Program Lead 14 ($378K–$441K, SF + NYC) focused on enterprise AI adoption programs, a Product Lifecycle Manager 15 ($250K–$300K), a Market Research Lead 16 ($234K–$260K), and an Art Director (remote US) for the Marketing creative team.

Anthropic added 27 listings to its Greenhouse board 17 between May 4–11, growing its total open-role count from roughly 393 to 420 — a 6.4% net addition in a single week. The postings span 8+ departments and carry a clear geographical expansion thread.

Two brand-new platform teams. On May 5, Anthropic posted founding Engineering Manager roles for two previously unannounced product groups. The Marketplace platform 18 ($405K–$485K) will build customer-facing discovery surfaces for enterprises and partner-facing self-serve tooling. The Passport platform 19 ($405K–$485K) is a centralized identity, verification, and trust-grant layer spanning Claude.ai, the API, and third-party integrations. Both are 0→1 roles: the founding EMs will hire their teams from scratch.

Government and public sector. A Staff+ Software Engineer, Public Sector 20 ($405K–$485K, SF or NYC) requires US citizenship and a preference for an active federal security clearance at Secret level or above. The job description frames the role as "bringing Claude to democracies worldwide" — Anthropic's most explicit public-sector recruiting signal to date.

Dublin compliance hub. A cluster of European roles points to a new compliance and trust & safety center in Ireland. Postings from May 4–6 include a Content Moderation Specialist, Privacy Governance Lead, IT Support Engineer, Policy Counsel EMEA, AI Compliance Officer, Corporate Compliance Officer, Sanctions & Export Controls Compliance Officer, and Commercial Counsel Infrastructure Security — all Dublin-based and several referencing GDPR frameworks.

London infrastructure engineering. Four Staff-level infrastructure roles posted May 6–7 in London: two Staff Software Engineers for Kubernetes Platform, one Staff Software Engineer for Node Infrastructure, and one Staff Infrastructure Engineer for Cluster Infrastructure. UK salary ranges are not legally required to be disclosed, and Anthropic omitted them.

Safety documentation. A Research Operations, External Artifacts 21 role ($260K–$310K, remote-friendly US) owns Anthropic's system card production process — the long-form technical documents published with each model release that describe capabilities, safety evaluations, and deployment rationale. It coordinates across Alignment, Safeguards, and the Frontier Red Team.

Across all US-based Anthropic postings this week, compensation ranged from $170K (Contracts Manager) to $550K OTE (Manager, Strategic Sales). Staff and senior engineering roles cluster at $320K–$485K.

xAI posted 10 new roles against its 253-total-open-position board 22 during the May 4–11 window. Seven of the ten are operations roles tied directly to xAI's data center expansion in Memphis, Tennessee.

One software engineering role — Member of Technical Staff, Cloud Infrastructure (Palo Alto) — appeared against that backdrop of physical buildout. The only other non-data-center postings were a Corporate Counsel and a Safety Trainer for monetization. Salary ranges are not disclosed on xAI's Greenhouse listings. External reporting from May 2026 noted workforce reductions at xAI during the same period, though the new postings data does not confirm or contradict that directly.

metacareers.com does not display posting dates on job detail pages or in search results, so verifying which Meta roles are genuinely new this week requires relying on Google's freshness indexing. Only one Meta role was confirmed by Google's index as first appearing within the May 4–11 window.

Confirmed in-window: A Research Scientist, AI, Formal and Informal Reasoning 23 posted at FAIR Paris focuses on advancing LLM reasoning using reinforcement learning, synthetic data generation, and agentic techniques. The role requires a PhD with first-author publications at NeurIPS, ICML, or ICLR and familiarity with the Lean 4 theorem-proving language. No salary range is listed for France-based roles.

Meta Superintelligence Labs (MSL) — a new organizational structure announced in 2025 — now houses FAIR (Fundamental AI Research, Meta's core research division) alongside dedicated foundational modeling teams. Several currently open (undated) roles sit explicitly within MSL, including an AI Research Scientist on the Evaluations team ($154K–$217K, Menlo Park) and a Speech & Audio LLMs Research Scientist ($122K–$181K, Menlo Park). Because posting dates are unavailable for these roles, they are context for understanding Meta's AI research structure rather than confirmed additions this week.

Meta's US-based AI Research Scientist salary bands run from $122K (entry PhD, Menlo Park) up to $301K (Technical Leadership tier), lower than Anthropic's equivalent IC bands but with additional bonus and equity components.

Google DeepMind's Greenhouse board 24 posted a single role in the May 4–11 window: a UX Researcher, 12-month Fixed Term Contract 25 on the Inception AIUX (Artificial Intelligence User Experience) team, based in London. The Inception team builds "0→1 breakthrough experiences" at the frontier of generative AI through rapid prototyping and user research. No salary is disclosed.

DeepMind maintains 67 total open roles on Greenhouse. Its careers page notes that open positions are actively transitioning to Google Careers, where an additional 48 roles are listed — but Google Careers does not expose first-published dates in a way that allows weekly new-posting detection. The full scope of DeepMind's hiring activity is therefore wider than Greenhouse alone reflects.

Mistral AI posted 2 new roles on May 5, 2026, both in the Solutions team 26 — one in Paris and one in Abu Dhabi, UAE. Mistral appears to batch-post jobs weekly (the prior batches land on Apr 30, Apr 23, and similar intervals). The Solutions team is Mistral's largest hiring area, with approximately 40+ active roles globally. Job titles for the two May 5 postings could not be extracted from the Lever API response due to encoding issues; direct links to both: Paris role. Mistral has approximately 110 total open roles across 11 departments including Research, Engineering & Infra, and Business.

Cohere lists 127 open roles on its Ashby board 27, but Ashby does not surface posting dates or recency markers — no weekly new-posting count can be confirmed for Cohere. For context on the overall shape of Cohere's hiring: the Modeling department leads with approximately 20 roles (primarily Member of Technical Staff and Senior Research Engineer), followed by Agentic Platform (~16 roles, concentrated in Forward Deployed Engineers), and Applied-ML (~10). Roles explicitly titled "Sovereign AI" and "AI Delivery for Public Sector & Defence" point to a government-market strategy. Geographic reach spans Toronto, London, Paris, Seoul, Singapore, Dubai, Riyadh, and several other cities.

Anthropic is the most aggressive hirer right now. Twenty-seven new postings in seven days — roughly 27× DeepMind's weekly Greenhouse velocity and nearly 3× xAI's confirmed count — spans technical and non-technical functions in roughly equal measure. The simultaneous launch of two 0→1 platform teams (Marketplace, Passport) on the same day is the clearest product infrastructure signal of the week.

Safety and alignment hiring is accelerating at two labs simultaneously. OpenAI opened two safety research roles (Alignment Oversight, Recursive Self-Improvement Preparedness) while Anthropic posted a Research Operations role dedicated to system card production. These are not junior support roles; the OpenAI positions carry base ranges up to $445K.

OpenAI's ads cluster is real. Three distinct ads and revenue-operations roles posted within two days is a material signal that ChatGPT's advertising product is moving from idea to org-building.

xAI is building infrastructure, not software. Seven of its ten new roles this week are electrical and mechanical commissioning technicians, managers, and security supervisors for data center facilities in Memphis. One software engineering hire appeared. Whether this ratio reflects a deliberate buildout phase or a consequence of the reported workforce reductions in other departments is unclear from job postings alone.

Government and public-sector AI is becoming a dedicated function. Anthropic's US-citizenship-required Staff+ SWE, Public Sector and Cohere's sovereign AI and public-sector defence roles, alongside OpenAI's Partnerships Lead, Healthcare, point to a broader pattern: frontier labs are building dedicated government-facing teams rather than treating public-sector work as a sales-team side effort.