AI Labs Hiring Weekly #3 — OpenAI, Anthropic, DeepMind, Meta AI & xAI

Two Anthropic roles arrived freshly flagged as new on the Greenhouse board: Product Engineer, Computer Use and Data Operations Manager, Human Data — both carry published salary bands, adding to Anthropic's already dense RL/research hiring push. OpenAI's Codex team now has its own Research Engineer listing at $295K–$445K, a signal that the Codex hiring push is moving beyond post-launch sprint mode and into sustained staffing. Google DeepMind broke its usual salary silence on two postings this week — Research Engineer, Human Understanding ($174K–$252K) and Research Engineer, Frontier Safety Risk Assessment ($136K–$245K) — the first DeepMind US salary bands confirmed in this digest. xAI added MTS Post-Training & RL (Palo Alto) to its growing model-team roster.

Week-over-week salary changes: none. All previously reported bands at Anthropic ($500K–$850K for RL Engineering), xAI ($180K–$600K MTS Model Training), and OpenAI ($295K–$555K Post-Training RE/RS) are unchanged.

OpenAI (717 open roles as of this week) continues to hire primarily through Ashby at jobs.ashbyhq.com/openai; the older openai.com/careers/search page mirrors roles but Ashby is the authoritative source.

Research Engineer, Codex — San Francisco, hybrid (3 days/week). The Codex team owns code-generation and agentic-coding products end-to-end across research, engineering, and infrastructure. This role sits at the intersection of ML research and systems performance, with tasks ranging from training-run experiments to LLM inference optimization. Compensation: $295K–$445K + equity. 1

Research Engineer / Research Scientist, Post-Training — San Francisco, hybrid. The Post-Training team prepares pre-trained models for deployment in ChatGPT and the API through RL-focused research. This is one of OpenAI's widest salary windows, reflecting a range that spans research engineer to research scientist levels. Compensation: $295K–$555K + equity. 2

Software Engineer, ML Systems & Training Architecture — San Francisco, on-site (5 days/week, stricter than most OpenAI roles). This is under the Robotics team and focuses on maintaining training-framework health, debugging ML system failures across GPU clusters, and unblocking researchers. The 5-day in-office requirement distinguishes it from the typical hybrid policy. Compensation: $295K–$380K + equity. 3

Also active on Ashby: Research Engineer/Research Scientist, RL/Reasoning (o1/o3 team); Research Engineer, Speech & Realtime Models; Researcher, Pretraining Safety; Researcher, Connectors — Agent Post-Training. Salary bands for those roles were not accessible via static HTML this week.

Anthropic's Greenhouse board shows two roles with "New" badges this week, consistent with the hiring acceleration noted in the last issue. Both carry published salary bands.

Product Engineer, Computer Use 🆕 — San Francisco / Seattle. Works on Anthropic's computer-use API capability: full-stack development (web front-end + back-end), LLM API integration, browser/desktop automation tooling, and evaluation frameworks for agentic systems. Compensation: $300K–$320K + equity. 4

Data Operations Manager, Human Data 🆕 — San Francisco / New York. Manages human data operations feeding Claude's training pipelines. Compensation: $270K–$365K + equity. 5

Machine Learning Systems Engineer, RL Engineering — SF / NYC / Seattle. Core infrastructure for large-scale RL training: Python, distributed systems at scale, RLHF/finetuning pipelines. Requires at least 25% in-office time. This is the role with Anthropic's highest publicly disclosed band. Compensation: $500K–$850K + equity. 6

Full-Stack Software Engineer, Reinforcement Learning — SF / NYC. Builds infrastructure and tooling supporting Claude's RL team. Compensation: $300K–$405K + equity. 7

Remote-eligible roles (with travel requirement): Research Engineer, Environment Scaling; Research Engineer, Knowledge Team; Research Engineer/Scientist, Pre-training (US); Research Engineer, Machine Learning (RL Velocity); Research Engineer, Model Evaluations. Multiple RL and Pretraining openings also list London and Zürich locations.

Other active research postings: Research Engineer Interpretability, Research Engineer Pretraining Scaling, Research Engineer Cybersecurity RL, Research Engineer AI Observability, ML/Research Engineer Safeguards, Performance Engineer (Inference Systems), Engineering Manager GPU, Engineering Manager Inference.

Research Engineer, Human Understanding — Los Angeles / Mountain View. Works on Gemini multimodal reasoning for information literacy in the Frontier AI division. Tech stack: Python, LLM training and inference engineering. Compensation: $174K–$252K + bonus + equity. First DeepMind role with a published US salary band in this digest. 8

Research Engineer, Frontier Safety Risk Assessment — London / NYC / San Francisco. LLM safety evaluation research embedded in the GenAI division. Python core, LLM/foundation model background preferred. Compensation: $136K–$245K + bonus + equity. 9

Research Scientist, Multimodal Alignment, Safety, and Fairness — Kirkland / Mountain View / NYC. Pushes state of the art in VLM alignment and safety. Tech stack: Python, JAX/Flax, RL-based LLM finetuning, multimodal learning. Compensation: $147K–$211K + bonus + equity. 10

Research Scientist, Reinforcement Learning — London. Frontier AI division. Salary not disclosed. 11

Research Engineer, Multimodal Generative AI (Image/Video) — Kirkland / Seattle (GenAI). Salary not disclosed. 12

Also open: Product Manager, Information Quality, Frontier AI (Mountain View); Technical Program Manager, Agents Innovation (London, 12-month FTC); Technical Program Manager, Antigravity and Antigravity Modeling & Evals (Mountain View); Research Scientist Gemini Safety (Mountain View + Zürich); Senior Data Scientist, Gemini App (Zürich); Director Engineering, Create — GeminiApp (Mountain View). Salary not publicly disclosed on most DeepMind postings.

Meta's careers portal requires JavaScript to load listings, so role discovery this week relied on individual posting URLs found via search. Salary confirmation was obtained from posting detail pages where static HTML rendered.

Meta continues to hire under the Meta Superintelligence Labs (MSL) brand alongside the existing FAIR brand for fundamental research.

AI Research Scientist, VLM (Vision Language Models) — Bellevue, WA + 1 location. Pushes state of the art in multimodal generative AI for AI assistants. PhD in AI/CS required. Tech stack: Python, PyTorch. Compensation: $184K–$257K + bonus + equity. 13

Research Engineer, MSL FAIR Foundations — Menlo Park + 1 location. Builds and evaluates benchmarks for MSL's most advanced models across text, vision, audio. Tech stack: Python, PyTorch, large-scale distributed systems, RL environments. Compensation: $183K–$257K + bonus + equity. 14

Also listed but salary not confirmed via static page this week: AI Research Engineer, FAIR Chemistry; AI Research Scientist, SysML — FAIR; Research Engineer, MRS AI; Software Engineer, Systems ML. Meta's total listed AI roles are not accessible in aggregate without JavaScript rendering.

xAI's full roster is at x.ai/careers/open-roles and spans 228+ positions. The Greenhouse board at job-boards.greenhouse.io/xai provides the structured listing.

MTS — Model Training — Austin / NYC / Palo Alto / Seattle. Focuses on critical modeling challenges at any given time; first project disclosed before offer. Compensation: $180K–$600K + equity. 15

MTS — Post-Training and RL 🆕 — Palo Alto. RLHF/DPO, reward modeling, RL for reasoning and truthfulness. Compensation: $180K–$600K + equity. 16

MTS — Pre-Training — Palo Alto. Trillion-parameter training at scale, distributed multi-GPU training, model-hardware co-design. Compensation: $180K–$440K + equity. 17

MTS — Multimodal Understanding — Palo Alto. Salary not disclosed on this posting. 18

MTS — Mid-training — Palo Alto. Salary not disclosed. 19

Exceptional Software Engineer — Austin / NYC / Palo Alto / Seattle. The generalist high-bar role for product and technical challenges. Compensation: $180K–$440K + equity. 20

Backend Engineer — API — London. xAI's London office continues to expand; the full London roster includes MTS Model Training, MTS Reasoning, Backend Engineer, SRE, MTS Sandbox Service, and Exceptional Software Engineer. Salary not disclosed on London postings. 21

xAI does not use external recruiters for initial screening; all applications go through the Greenhouse board directly.