Active Roles (7 Labs)
702+
Labs Covered
7
Salary Transparent
4 / 7
Remote-Friendly
3 / 7
AI Lab Job Digest — Week of May 27, 2026
A structured weekly digest of open SWE, ML/Research, and PM roles at seven top AI labs (OpenAI, Anthropic, Google DeepMind, Meta AI, xAI, Mistral AI, Cohere), covering 20+ role listings with a cross-company compensation comparison table, per-role field cards (title, team, location, salary, stack, apply link).
Seven top AI labs had a combined 702+ active open roles on their careers pages as of May 27, 2026. 1 2 3 This digest surfaces the most actionable SWE, ML/Research, and PM openings across OpenAI, Anthropic, Google DeepMind, Meta AI, xAI, Mistral AI, and Cohere — with compensation, location, and stack details where available.
正在加载统计卡片…
One structural caveat: none of these labs display a "date posted" field on individual listings, making it impossible to confirm which roles appeared in the last four weeks versus earlier. All roles below were verified as active on May 27, 2026. Future rounds of this digest will diff job IDs week-over-week to surface genuinely new postings.
Compensation at a glance
Before diving per-company, here's how the salary landscape compares across labs for SWE and ML roles. Numbers are base salary unless noted; TC (total compensation) figures are from Levels.fyi and include equity and bonus.
| Lab | Disclosed base range (SWE/ML) | Levels.fyi TC median | Salary transparency |
|---|---|---|---|
| OpenAI | $185K–$555K (role-dependent) | $608K (SWE L4) | ✅ Most postings |
| Anthropic | $350K–$850K (ML roles) | $674K (SWE) | ✅ ML/Infra postings |
| Google DeepMind | $166K–$252K (US only) | $327K (Google SWE) | ⚠️ US postings only |
| Meta AI | $154K–$217K (1 of 11 roles) | $430K (Meta SWE) | ❌ Most pages inaccessible |
| xAI | $180K–$600K (MTS/SWE) | $640K (SWE) | ✅ Key postings |
| Mistral AI | Not disclosed | Not available | ❌ No public figures |
| Cohere | Not disclosed | $151K (SWE TC) | ❌ No public figures |
The pay gap between Anthropic's top ML band ($850K base) and DeepMind's US research roles ($252K base) reflects entirely different compensation models — Anthropic front-loads base salary while Google packages most upside in equity and RSUs under the broader Google compensation grid.
OpenAI
OpenAI is an AI research and deployment company headquartered in San Francisco; it operates under a nonprofit Foundation and a for-profit subsidiary structured as a public benefit corporation. Main products include ChatGPT, GPT-series models, the Codex coding agent, and Sora. 11 As of May 27, the careers portal lists 702 active roles across all departments, managed via Ashby (OpenAI migrated from Greenhouse;
boards.greenhouse.io/openai now returns 404). 1Highlighted openings
Research Engineer (pipeline) — Research team 12
- Location: San Francisco, CA (onsite)
- Base salary: $250K–$445K + equity
- Requirements: Strong programming skills; large distributed systems experience; high-performance deep learning implementations preferred
- Apply: jobs.ashbyhq.com/openai
Research Scientist (pipeline) — Research team 13
- Location: San Francisco, CA (onsite)
- Base salary: $250K–$445K + equity
- Requirements: First-author ML publication track record or equivalent project portfolio; ability to drive a research agenda independently
- Apply: jobs.ashbyhq.com/openai
Backend Software Engineer, ChatGPT Engineering — Applied AI Engineering 14
- Location: San Francisco, New York, Seattle (onsite)
- Base salary: $185K–$385K
- Stack: Go, Python, Rust, C++; distributed systems; data storage and retrieval
- Apply: jobs.ashbyhq.com/openai
Software Engineer, Codex Core Agents — Codex Engineering 15
- Location: San Francisco, New York, Seattle, London (onsite)
- Base salary: $230K–$385K + equity
- Stack: Distributed systems, containers/sandboxing/virtualization, agent orchestration, LLMs, CI/CD, security engineering
- Apply: jobs.ashbyhq.com/openai
Machine Learning Engineer, Integrity — Applied AI Engineering 16
- Location: San Francisco, CA (onsite)
- Base salary: $266K–$555K + equity
- Stack: PyTorch, TensorFlow; transformer models; LLM training (distillation, SFT, policy optimization); abuse prevention
- Education: Master's or PhD in CS, ML, Data Science, or related field required
- Apply: jobs.ashbyhq.com/openai
Product Manager, API Agents — Product Management 17
- Location: San Francisco, CA (onsite; relocation assistance available)
- Base salary: $293K–$325K + equity
- Requirements: 5+ years PM experience; developer-facing products; API/SDK design intuition; experience with high-growth tech companies
- Apply: jobs.ashbyhq.com/openai
Also open across the same careers page: Research Engineer / Research Scientist roles in Post-Training, RL/Reasoning, and Alignment; mobile and frontend engineers on the ChatGPT team; and PM roles for Core Identity, Premium Subscriptions, and Safety Measurement. 1
Early-career note: OpenAI's Emerging Talent program covers the 0–3 years experience bracket across research, engineering, and product. A 6-month Residency is also available for engineers or researchers switching into AI. Talent community signup: jobs.ashbyhq.com/openai/form/openai-emt-talent-community. 18
Anthropic
Anthropic is an AI safety and research company based in San Francisco, incorporated as a public benefit corporation. It develops the Claude series of AI assistants and Claude Code (an agentic coding tool) built on a mission of making AI systems that are reliable, interpretable, and steerable. 19 Anthropic's Greenhouse board showed 388 active roles as of May 27. 2
Highlighted openings
Research Engineer, Machine Learning (Reinforcement Learning) — AI Research & Engineering 20
- Location: San Francisco or New York (hybrid; also London)
- Base salary: $500K–$850K
- Stack: Python; PyTorch / JAX / TensorFlow; industrial ML research experience; RL training pipelines
- Note: This team contributed directly to Claude Sonnet 4.5 and Opus 4.5's autonomy and coding capabilities
- Deadline: Rolling
Performance Engineer, Inference Systems — AI Research & Engineering (New) 21
- Location: San Francisco or New York (hybrid)
- Base salary: $350K–$850K
- Stack: Python performance engineering; SQL/pandas for data analysis; familiarity with GPU/TPU accelerator performance concepts
- Focus: Cross-dimensional optimization of inference systems across throughput, latency, reliability, and correctness
- Deadline: Rolling
Staff SWE, Claude Code — Engineering & Design - Product
- Location: San Francisco (hybrid)
- Base salary: Not listed (Levels.fyi median SWE TC at Anthropic: $674K) 5
- Stack: Systems-level development for an agentic coding tool; full-stack or backend depth
- Apply: boards.greenhouse.io/anthropic
- Deadline: Rolling
Product Manager, Claude Code — Product Management
- Location: San Francisco or Seattle (hybrid)
- Base salary: Not listed (Levels.fyi PM median at Anthropic: $545K) 5
- Deadline: Rolling
- Apply: boards.greenhouse.io/anthropic
Anthropic Fellows Program 2026 — open now for rolling admission 22
- Type: 4-month paid research fellowship (not a traditional job listing)
- Stipend: $3,850/week USD (≈$61,600 over 4 months) + ~$15,000/month in compute credits
- Locations: Berkeley, CA or London (shared office space); remote option for US/UK/Canada residents
- Work authorization: US, UK, or Canada work permit required; no visa sponsorship
- Tracks: AI Safety, AI Security, ML Systems & Performance, Reinforcement Learning, Economics & Policy
- Outcomes: 25–50% of prior fellows received full-time Anthropic offers; 80%+ published research
Remote and visa policy: Anthropic expects all staff to be in an office at least 25% of the time. 23 Most engineering roles offer H-1B and green card sponsorship, though not all positions guarantee it.
Google DeepMind
Google DeepMind is Alphabet's AI research lab, formed by merging Google Brain and DeepMind in 2023 under CEO Demis Hassabis. 24 It builds Gemini (Alphabet's flagship model family), AlphaFold, Veo (video generation), and robotics systems, among others. As of May 27, DeepMind had 59 open roles on Greenhouse and 66 on Google Careers. 3 25 US-based roles carry disclosed salary bands; roles in London, Zürich, and Tokyo do not.
Highlighted openings
Research Engineer, Human Understanding (L5) — Frontier AI 26
- Location: Los Angeles or Mountain View, CA
- Base salary: $174K–$252K + bonus + equity + benefits
- Requirements: PhD in CS or ML + 3 years experience; multimodal model development (audio/speech-visual); Python + JAX; top-venue publication record (NeurIPS/ICML/ICLR/CVPR)
- Focus: Foundational multimodal AI for human understanding, generation, and deepfake detection
Research Engineer, Multimodal Generative AI (Image/Video) — GenAI "Nano Banana" team 27
- Location: Kirkland or Seattle, WA
- Base salary: $166K–$244K + bonus + equity + benefits
- Requirements: PhD in CS/AI/ML/CV; generative AI experience (diffusion models, autoregressive models); Python + JAX/TensorFlow/PyTorch; top-venue publication record
- Apply: job-boards.greenhouse.io/deepmind
Research Scientist, Reinforcement Learning — RL team (London) 28
- Location: London, UK
- Base salary: Not listed (UK posting; no pay transparency requirement)
- Requirements: RL research publication record; strong implementation skills (JAX/Flax or PyTorch); end-to-end experimental experience; PhD in ML preferred
- Note: The RL team has built DQN, AlphaGo, AlphaZero, MuZero, AlphaStar, AlphaProof, and Gemini
- Apply: job-boards.greenhouse.io/deepmind
Software Engineer, Numerics — GenAI 29
- Location: Mountain View, CA
- Base salary: Not listed
- Requirements: PhD in CS + 2 years experience; deep expertise in numerical precision, quantization, sparsification; low-precision and sparse model deployment
- Note: This role directly influences the architecture roadmap of Google's custom AI chips (TPUs)
- Apply: job-boards.greenhouse.io/deepmind
Director of Product Management, Robotics 30
- Location: Mountain View, CA
- Base salary: $272K–$383K + bonus + benefits
- Requirements: Master's in Robotics or CS; 10+ years PM experience with at least 5 years leading AI/foundation model/robotics products; 5+ years managing senior PMs
- Apply: job-boards.greenhouse.io/deepmind
Two other PM roles (Group PM, Agent Security and PM, Information Quality) had application deadlines that passed in April 2026 and are excluded.
Meta AI
Meta is a social technology company (Facebook, Instagram, WhatsApp, Threads) with two AI research divisions: FAIR (Fundamental AI Research) and Meta Superintelligence Labs (MSL). 31 Meta is investing $135 billion in capex in 2026, primarily for AI infrastructure. Most of Meta's careers pages at metacareers.com use JavaScript rendering that blocked automated data collection this week — the detail pages for approximately 10 of 11 identified AI roles could not be retrieved. The roles below are listed at the level of detail that was recoverable.
Identified openings (⚠️ limited detail — detail pages inaccessible)
- Tech Lead, AI Research Scientist (Robotics) — Meta Superintelligence Labs — approx. posted May 19 — metacareers.com
- Research Scientist — FAIR — posted May 13 — metacareers.com
- Research Scientist, Artificial Intelligence — MSL — approx. posted May 21 — Requirements include peer-reviewed AI publications or open-source ML contributions — metacareers.com
- Research Engineer, Evaluations (Tech Leadership) — approx. posted May 23 — Requirements: 5+ years ML engineering; building AI benchmarks across text/vision/audio — metacareers.com
- Technical Program Manager, FAIR — posted May 27 — metacareers.com
One fully detailed role:
AI Research Scientist — Generative AI for Materials Discovery — Reality Labs Research 32
- Location: Redmond, WA
- Base salary: $154K–$217K + bonus + equity + benefits
- Requirements: Bachelor's + PhD in ML/computational chemistry/materials science/physics; 3+ years generative modeling research (diffusion, flow matching, VAEs, autoregressive models); Python + PyTorch/JAX; top-venue publications (NeurIPS, ICML, ICLR, JACS)
- Focus: Applying generative AI to design novel compounds for AR/VR devices and robotics materials
Meta's hiring across broader AI is constrained by an ongoing restructuring (~700 job cuts reported in 2026), though AI investment remains a stated priority.
xAI
xAI was founded by Elon Musk in 2023 with the mission to "accelerate human scientific discovery." 33 Its core product is the Grok reasoning model (currently at version 4.3); it also operates the Colossus supercomputer cluster (200,000 GPUs, built in 122 days). xAI secured a $200 million U.S. Department of Defense contract in May 2026 and is actively expanding. 33 All roles are onsite; xAI operates a flat, hands-on culture with no traditional PM titles — product responsibilities are embedded in "Member of Technical Staff" (MTS) roles.
Highlighted openings
Exceptional Software Engineer 34
- Location: Austin TX, New York NY, Palo Alto CA, Seattle WA (onsite)
- Base salary: $180K–$440K + equity
- Requirements: Exceptional engineering skills; belief in truth-seeking AI as the most important problem; no specific stack or language constraint listed
- Note: Unusually sparse JD — candidates receive clarity on their first project before an offer is made
Member of Technical Staff — Model Training 35
- Location: Austin TX, New York NY, Palo Alto CA, Seattle WA + London UK (onsite)
- Base salary: $180K–$600K + equity
- Requirements: "Power user of AI models"; experience training models used by millions is a plus but not required; no formal ML background mandated
Member of Technical Staff — Pre-Training — Apply
- Location: Palo Alto, CA (onsite)
- Base salary: Not listed on careers page (reference range $180K–$600K from comparable MTS roles)
- Focus: Large-scale language model pretraining
Member of Technical Staff — Post-Training and RL — Apply
- Location: Palo Alto, CA (onsite)
- Focus: Post-training alignment techniques (RLHF, DPO) and reinforcement learning
Levels.fyi puts the xAI SWE median total compensation at $640K (as of May 26, 2026), driven by substantial equity in a company that raised at high valuations. 8
Mistral AI
Mistral AI is a French AI startup founded in April 2023 by Arthur Mensch, Guillaume Lample, and Timothée Lacroix (former Google DeepMind and Meta researchers), headquartered in Paris. 36 Products include Mistral Small 3 (open-source models), the Le Chat AI assistant, and the La Plateforme / AI Studio enterprise platform. None of Mistral's job postings list salary figures; all describe compensation as "competitive salary and equity." 37
Highlighted openings
Software Engineer, Backend — Engineering & Infra 38
- Location: Paris, France (onsite, 3+ days/week) or New York, NY (hybrid) — separate listings
- Salary: Not listed
- Stack: Python (primary); JavaScript/TypeScript, C#, Golang also accepted; APIs, databases, caching, distributed architectures; bonus: Docker, CI/CD, Kubernetes, Helm, Terraform, Prometheus, Grafana
- Requirements: CS degree or equivalent; ownership mindset across the full feature lifecycle
- Interview process: 5–6 rounds (intro, live coding, system design, hiring manager, culture-fit, references)
- Apply (Paris): jobs.lever.co/mistral
AI Scientist — Research (Paris/London) 39
- Location: Paris or London (hybrid)
- Salary: Not listed
- Stack: Python (or Rust/Go/Java); PyTorch/JAX or Ray/Kubernetes for distributed systems; distributed transformer training experience preferred; strong publication record valued
- Focus: Novel LLM methods covering reasoning, code, agents, and multimodal (text/image/speech)
Research Engineer, Machine Learning — Research (Paris/London/Zürich/Warsaw) 40
- Location: Paris, London, Zürich, or Warsaw (hybrid)
- Salary: Not listed
- Stack: PyTorch/JAX/TensorFlow; distributed training (DeepSpeed, FSDP, SLURM, Kubernetes); 4+ years large-scale ML codebase experience; Master's or PhD preferred
- Tracks: Platform RE (shared training framework) or Embedded RE (aligned to a specific research squad)
Product Manager, AI Studio — Product (Paris) 41
- Location: Paris, France (onsite, 3+ days/week)
- Salary: Not listed
- Requirements: 5+ years PM; hands-on generative AI/LLM product experience; ability to prototype/build; cross-functional influence without direct authority
- Focus: AI Studio enterprise platform covering observability, API/developer experience, workflows, and connectors
Mistral also lists a New Grad SWE opening in Paris and a PM, Cloud Partnerships role in New York. 37
Cohere
Cohere was founded in 2019 in Toronto by Aidan Gomez (CEO and co-author of the "Attention Is All You Need" Transformer paper), Nick Frosst, and Ivan Zhang. 42 It focuses on enterprise AI, offering the Command A+ model, the North agentic enterprise platform, and Embed/Rerank semantic search tools. Cohere completed a $600 million Series E in April 2026 at a $6.8 billion valuation. 42 No compensation figures appear in Cohere job postings; Levels.fyi reports a SWE median TC of $151K (Canada-weighted). 9
Highlighted openings
(Listed via Welcome to the Jungle — Cohere's own Ashby portal is a JS SPA that was inaccessible this week. Apply links below go through the WTJ platform.)
Senior/Staff Full-Stack Engineer — Software Engineering 43
- Salary: Not listed (Levels.fyi SEM median at Cohere: $333K TC) 9
- Apply: welcometothejungle.com
Staff Software Engineer, Inference Infrastructure 43
- Focus: Inference infrastructure at scale
- Salary: Not listed
- Apply: welcometothejungle.com
Senior Member of Technical Staff (Multimodal AI) — Research 43
- Focus: Multimodal AI research and development
- Apply: welcometothejungle.com
Product Manager (Agent Harness & Modelling) — Product 43
- Salary: Not listed (Levels.fyi PM median: $189K TC) 9
- Apply: welcometothejungle.com
Software Engineer Intern (Fall/Winter 2026) — Remote, Canada 44
- Duration: 3–6 months full-time; open to current post-secondary students
- Location: Remote, Canada
- Scope: Frontend, backend, full-stack, or infrastructure — interns ship production code
- Apply: via Built In or Cohere's Ashby portal
Welcome to the Jungle lists 72 total open roles at Cohere. 43
Employer quick-reference
正在加载统计卡片…
| Company | HQ | ATS | SWE/ML/PM? | Salary transparency | Remote policy |
|---|---|---|---|---|---|
| OpenAI | San Francisco | Ashby | ✅ All three | ✅ Most listings | Onsite (SF/NYC/SEA/London) |
| Anthropic | San Francisco | Greenhouse | ✅ All three | ✅ ML/Infra; PM not listed | 25% office minimum |
| Google DeepMind | London / Mountain View | Greenhouse + Google Careers | ✅ All three | ✅ US only | Varies by role |
| Meta AI | Menlo Park | Meta Careers | ✅ All three | ❌ Pages inaccessible | Varies |
| xAI | Palo Alto | Greenhouse | SWE + ML (no standalone PM) | ✅ Key postings | Onsite required |
| Mistral AI | Paris | Lever | ✅ All three | ❌ None disclosed | Hybrid (Paris-centric) |
| Cohere | Toronto | Ashby (via WTJ) | ✅ All three | ❌ None disclosed | Remote-friendly |
Cover image: AI-generated illustration.
参考来源
- 1Careers at OpenAI
- 2Jobs at Anthropic — Greenhouse
- 3Jobs at DeepMind — Greenhouse
- 4OpenAI Salaries — Levels.fyi
- 5Anthropic Salaries — Levels.fyi
- 6Google SWE Salaries — Levels.fyi
- 7Meta SWE Salaries — Levels.fyi
- 8xAI Salaries — Levels.fyi
- 9Cohere Salaries — Levels.fyi
- 10Pixabay — AI workspace photo
- 11About
- 12Research Engineer
- 13Research Scientist
- 14Backend Software Engineer, ChatGPT Engineering
- 15Software Engineer, Codex Core Agents
- 16Machine Learning Engineer, Integrity
- 17Product Manager, API Agents
- 18Emerging Talent
- 19Making AI systems you can rely on — Anthropic
- 20Research Engineer, ML (RL) — Anthropic
- 21Performance Engineer, Inference Systems — Anthropic
- 22Anthropic Fellows Program
- 23Anthropic Careers
- 24About Google DeepMind
- 25Google Careers — DeepMind roles
- 26Research Engineer, Human Understanding — DeepMind
- 27Research Engineer, Multimodal GenAI — DeepMind
- 28Research Scientist, Reinforcement Learning — DeepMind
- 29Software Engineer, Numerics — DeepMind
- 30Director of Product Management, Robotics — DeepMind
- 31About Meta
- 32AI Research Scientist — Materials Discovery — Meta
- 33Company — xAI
- 34Exceptional Software Engineer — xAI
- 35MTS, Model Training — xAI
- 36About us
- 37Mistral AI — Lever Job Board
- 38Software Engineer, Backend (Paris) — Mistral AI
- 39AI Scientist — Mistral AI
- 40Research Engineer, ML — Mistral AI
- 41Product Manager, AI Studio — Mistral AI
- 42About Our Company
- 43Cohere Jobs — Welcome to the Jungle
- 44SWE Intern — Cohere
围绕这条内容继续补充观点或上下文。