AI Jobs › Hiring brief archive

AI Hiring Brief · 2026-09-29

32 frontier AI labs · 57 roles added in 24h · agihunt.info/en/jobs

Today at a glance

  • Broad-based hiring day: across the 32 labs tracked, 57 new roles opened against 42 closures in the last 24 hours — one of the more active days recently, with Anthropic (+14/-8), Google DeepMind (+10/-3) and OpenAI (+7/-6) driving most of the volume.
  • DeepMind concentrates on safety: five of DeepMind's ten new roles today are explicitly safety- or security-tagged — Gemini Security Technical Program Manager, Research Scientist/Engineer for Autonomous Security, Introspection Probe Research Engineer (Gemini Security), Senior Technical Program Manager for AGI Safety and Alignment, and Senior Product Manager for Gemini App Youth Safety.
  • Anthropic pushes into new sales geographies: new postings in Mumbai, Sydney, Tokyo, London and Zürich span Enterprise Account Executive, Head of Customer Success APAC, Recruiter GTM and Applied AI Engineer/Architect roles, alongside infra hires in Cloud Inference and Scaling.
  • Data-center buildout keeps showing up: OpenAI's Data Center Hardware Quality & Reliability Engineer, Meta's Gigawatt Site Project Director for Data Centers, and SpaceXAI's Lead Technician (Power Generation) in Memphis all point to physical infrastructure hiring running in parallel with model work.
  • NVIDIA and Apple AIML net negative today: both closed more roles than they opened (NVIDIA 4 added/7 closed, Apple 2 added/5 closed), even as their 7-day totals stay firmly positive (+24 and +17 respectively).

Company moves

OpenAI (831 open, +7/-6 today, +70 over 7 days) added a Software Engineer, DevOps and a Data Center Hardware Quality & Reliability Engineer in Scaling, plus a Strategic Delivery Lead, Cyber under OpenAI for Gov in Washington, DC and an HR Business Partner, EMEA in Paris — infrastructure buildout running alongside government and international people-ops growth.

Anthropic (628 open, +14/-8 today, +57 over 7 days) posted the largest single-day net addition of any lab. New Staff/Sr. Software Engineer roles in Cloud Inference, Cloud Inference Launch Engineering and Scaling show continued infra investment, while Enterprise Account Executive (Mumbai), Head of Customer Success, APAC (Sydney), Recruiter, GTM (Tokyo) and several Applied AI Engineer/Architect roles in London mark an aggressive international commercial push.

Google DeepMind (谷歌 DeepMind) (80 open, +10/-3 today, +33 over 7 days) leaned heavily into safety and security hiring today, with five of its ten new roles tied to Gemini Security, autonomous-security research, AGI safety/alignment, and youth safety — alongside research roles in Agentic Post Training and Frontier Reasoning.

Meta AI (191 open, +5/-1 today, +33 over 7 days) opened a Gigawatt Site Project Director, Data Centers and a Software Engineer, Systems ML, plus a Security Engineer, Applied AI in London — infra and security growth running together.

NVIDIA (英伟达) (545 open, +4/-7 today, +24 over 7 days) closed more roles than it opened today despite steady weekly growth; new postings skew toward research and hardware, including a PhD Research Intern, Autonomous Systems and Physical AI Research (Zurich) and a System Optic PHY Engineer (Denmark).

Apple AIML (苹果 AIML) (291 open, +2/-5 today, +17 over 7 days) also ran net negative today, adding a Senior Data Scientist, Experimentation & Causal Inference (Cupertino) and a Software Development Engineer in Test, Siri AI QE (Yokohama).

SpaceXAI (289 open, +3/-0 today, +22 over 7 days) added infrastructure and security roles without closing any — Lead Technician (Power Generation) for its Memphis-area data center, an Infrastructure Security Engineer split between Dublin and London, and a Senior BSA/AML Investigator.

Cursor (128 open, +2/-0 today, +8 over 7 days) opened a Forward Deployed Engineer and a Regional Director, Forward Deployed Engineering, both for India — a clear signal of India market expansion.

Cohere (145 open, +1/-3 today, +8 over 7 days) ran net negative today, adding a single Solutions Architect role in Paris.

Cerebras (112 open, +2/-1 today, +6 over 7 days) added an ML Algorithm Mapping and Performance Engineer and an ML Runtime and Kernel Engineer, both on the Core ML team, in Sunnyvale and Toronto.

StepFun (阶跃星辰) (231 open, +1/-0 today, +5 over 7 days) added a 开放平台风控产品经理 (Open Platform Risk Control Product Manager) in Xuhui, Shanghai.

Meituan LongCat (美团·龙猫) (67 open, +0/-3 today, flat over 7 days) closed three roles today with no new postings — the largest same-day contraction among the Chinese labs.

Baidu AI (百度 AI) (475 open, +2/-4 today, +12 over 7 days) ran net negative today despite steady weekly growth, adding a 打车车点算法工程师 (ride-hailing dispatch algorithm engineer) and a 服务商质量管理 (service-provider quality management) role, both in Beijing.

ByteDance Doubao/Seed (字节跳动豆包/Seed) (893 open, +2/-1 today, +17 over 7 days) — the largest open count of any lab tracked — added two product roles: AI Agent产品经理-飞书&豆包工作 (AI Agent Product Manager, spanning Feishu and Doubao) and 大模型策略产品经理-豆包 (LLM Strategy Product Manager, Doubao), split between Beijing and Shanghai.

MiniMax (Hailuo) (MiniMax 海螺) (187 open, +2/-0 today, +3 over 7 days) added 全球 AI 商业化策略与生态合作 (Global AI Commercialization Strategy & Ecosystem Partnerships), spanning Shanghai, Hong Kong, Beijing, Singapore and San Francisco, plus a 高级税务经理 (Senior Tax Manager) in Shanghai.

The remaining 17 companies — Microsoft AI (微软 AI), Perplexity, Mistral AI, Lovable, Together AI, Scale AI, Pika Labs, Aleph Alpha, Genmo, Moonshot AI (月之暗面 Kimi), DeepSeek (深度求索), Zhipu AI (智谱 AI), ModelBest (面壁智能), Tencent Hunyuan (腾讯混元), 01.AI (Yi) (零一万物), SenseTime SenseNova (商汤日日新) and Alibaba Tongyi Qwen (阿里通义千问) — showed no change in open postings today.

What it signals

The infrastructure arms race remains the dominant thread in the 7-day data. OpenAI (+70), Microsoft AI (+58), Anthropic (+57), DeepMind (+33) and Meta AI (+33) are all adding at a fast clip, and a good share of today's new roles are physical rather than purely research-oriented: data-center hardware reliability at OpenAI, a gigawatt-scale site director at Meta, and power-generation and security technicians at SpaceXAI. NVIDIA's weekly total (+24) sits in the same range even though today itself was net negative, suggesting the closures are churn within a still-growing requisition base rather than a pullback.

DeepMind's cluster of safety and security roles today is unusually concentrated — half of its ten new postings touch security or alignment work, spanning red-team-adjacent research (Autonomous Security, Introspection Probe Research Engineer) and governance (AGI Safety and Alignment TPM, Youth Safety PM). That breadth, appearing in a single day rather than spread across weeks, points to safety headcount being built out as a distinct function rather than folded into general research hiring.

On the commercial side, Anthropic and Cursor are both visibly extending sales and deployment reach into new regions — Anthropic into India, Australia, Japan and the UK; Cursor specifically into India with both an individual-contributor and a regional-director role. MiniMax's Global AI Commercialization Strategy posting, spread across five cities including Singapore and San Francisco, is the clearest sign among the Chinese labs of a similar push toward international monetization, though it remains an isolated data point rather than a pattern across the group.

The Chinese labs overall are moving at a noticeably slower and more selective pace this week. ByteDance Doubao/Seed still carries the single largest open count of any lab tracked (893) and added a healthy +17 over seven days, but today's only two additions were product-management roles rather than research or infrastructure. Baidu AI and Meituan LongCat both ran net negative today, and half of the 13 Chinese labs tracked — DeepSeek, Zhipu AI, Moonshot AI, 01.AI, SenseTime SenseNova and Alibaba Tongyi Qwen — showed zero movement over the past week. That contrasts with the steady, multi-front expansion visible across the global labs, where infrastructure, safety, and international go-to-market hiring are all advancing at once.

← 2026-09-28 · All editions · 2026-09-30 →

All open roles at 32 frontier AI labs →