How to Get a Job at Anthropic: Hiring Process, Interviews and Salaries
Learn how to apply to Anthropic, what to expect in interviews, when candidates can use AI, and what salary ranges Anthropic publishes by role.

Facts checked on November 13–14, 2026.
LiftmyCV is not affiliated with Anthropic.
You want to work on frontier AI, Anthropic is near the top of your list, and the bar looks high from the outside. This guide explains how to get a job at Anthropic using what the company itself publishes. That means its careers page and its guidance on using AI when you apply. It also means the job postings live on its jobs board. On November 13, 2026, I opened 12 current Anthropic job postings, recorded their salary ranges and application questions, and paired them with open-role data from LiftmyCV's job database. Where Anthropic doesn't publish something, I say so. Where a detail comes from another company, I name it.
At a glance
- Open roles: 613 Anthropic roles in LiftmyCV's job database on November 14, 2026, most of them in engineering and sales.
- The AI rule: write your application yourself first, then you may use Claude to refine it. No AI in take-homes or live interviews unless Anthropic says otherwise.
- Pay: in the 10 postings I checked, US ranges ran from $220,000 to $850,000 a year. Sales figures are on-target earnings.
- Ways in: direct applications and the Anthropic Fellows Program. Anthropic says it doesn't currently offer internships.
- What helps: a specific, self-written answer to "Why do you want to work at Anthropic?", which the application form says Anthropic values highly.
What Anthropic looks for
Anthropic is an AI safety and research company and the maker of Claude. Every posting I opened starts with the same line: "Anthropic's mission is to create reliable, interpretable, and steerable AI systems."
Its careers page puts ability ahead of pedigree: "We care about what you can do, not where you learned to do it." Full-time postings ask for at least a bachelor's degree in a related field or equivalent experience. So the formal bar leaves room for people who learned on the job.
The same page lists the principles staff work by. They include "Do the simple thing that works," "Be good to our users" and "Put the mission first." Read them before you apply, because the forms ask about values directly. The Alignment research posting, for example, asks you to describe in one paragraph something meaningful you've done in line with your values.
Research is only part of the picture. On November 13, 2026, I found open postings in these functions:
- Research: Research Scientist, Interpretability (San Francisco)
- Engineering: Software Engineer, Claude Code (New York City, San Francisco, Seattle)
- Product: Product Manager, Revenue Platform
- Policy: Product Policy Manager, User Well-being
- Communications: Policy Communications Manager
- Go-to-market: Strategic Account Executive, Tech, and Applied AI Engineer (London)
Anthropic doesn't publish an ideal-candidate profile or a culture-fit checklist. What the postings do show is how much weight the written application carries. The application forms I read ask why you want to work at Anthropic. The Alignment form also asks which teams interest you and which three projects you'd want to work on.
If you're coming from big tech, our guide to how FAANG hiring works is a useful comparison. Anthropic's forms ask for far more writing up front.
Where Anthropic is hiring right now
On November 14, 2026, LiftmyCV's job database listed 613 open Anthropic roles, most of them in engineering and sales and based in San Francisco or New York City.
| Team (grouped from job titles) | Open roles | Share |
|---|---|---|
| Engineering | 174 | 28.4% |
| Sales and go-to-market | 160 | 26.1% |
| Research | 72 | 11.7% |
| Security | 71 | 11.6% |
| Finance | 27 | 4.4% |
| Product and design | 21 | 3.4% |
| Policy and legal | 15 | 2.4% |
| Operations | 12 | 2.0% |
| Marketing and communications | 10 | 1.6% |
| People and recruiting | 4 | 0.7% |
| Other | 47 | 7.7% |
Engineering and sales together make up more than half of open roles. Research accounts for 72 roles, about 12%, roughly the same as security. Our database has no team field, so we grouped roles by the words in their job titles.
| Location as listed in the posting | Open roles | Share |
|---|---|---|
| San Francisco or New York City | 177 | 28.9% |
| San Francisco, New York City or Seattle | 72 | 11.7% |
| San Francisco only | 68 | 11.1% |
| San Francisco or Seattle | 29 | 4.7% |
| London | 22 | 3.6% |
The table shows the top 5 location listings, which cover 368 of the 613 roles.
San Francisco appears in four of the five most common location listings. In the same data, 527 roles (86%) are tagged on-site, 61 (10%) remote and 25 (4%) hybrid. Those tags reflect how each posting is listed, not Anthropic's office policy, which the next subsection covers.
The 12 postings I opened on November 13 list offices in San Francisco, New York City, Seattle, Washington, DC, and London. Many roles name more than one city. A few are marked "Remote-Friendly (Travel Required)."
All 12 showed a salary range in the posting text, and several say applications are reviewed on a rolling basis with no deadline. LiftmyCV's data holds a structured salary range for 253 of the 613 roles (41%). Others may still state one in the text, since a range isn't always captured as a separate field. The median open role was first published on September 10, 2026, about nine weeks before our data pull. So a posting that has been up for months can still be worth applying to.
Offices and in-office expectations
Anthropic's postings say Anthropic currently expects all staff to be in one of its offices at least 25% of the time, and some roles require more. The application form asks whether you're open to that.
Remote exceptions exist but are narrow. The Interpretability research scientist posting is based in San Francisco and says Anthropic will consider exceptional candidates for remote work on a case-by-case basis.
Visa sponsorship
Anthropic's careers FAQ says it sponsors visas and green cards, though some operations roles are harder to support. Postings add that it can't sponsor every role and every candidate. The form asks whether you need sponsorship. Answer accurately, then ask your recruiter about your specific case.
To see Anthropic-style roles at other labs too, browse AI jobs hiring now on LiftmyCV.
How Anthropic's hiring process works
The Anthropic interview process starts with a written application. For technical roles it then usually moves through a recruiter call, a coding assessment and a final round of several interviews. Anthropic's careers page describes interviews in general terms only, so the stage details below come from interviewing.io's Anthropic guide, an interview-prep company, and are labeled that way.
- Written application. The form on the Software Engineer, Claude Code posting asks "Why do you want to work at Anthropic?" and says great answers are often 200–400 words. Research roles add team-specific questions.
- Recruiter call. interviewing.io, which says its information comes from conversations with Anthropic engineers, describes a 30-minute call.
- Coding challenge for technical roles. interviewing.io reports 60–90 minutes and says referred candidates may skip it.
- Final round. interviewing.io describes 4–5 hours for software engineers, usually five sessions: hiring manager, coding, system design, a second role-specific coding round and company values.
For technical interviews, the careers page says you can look things up but should be comfortable with basic syntax and standard libraries. Every process includes questions about your experience and motivation, plus time for your own questions.
How long it takes
Anthropic doesn't publish a timeline. interviewing.io puts the full process at about 3–4 weeks. Exponent says most candidates report 1 to 4 months from application to offer. Ask your recruiter on the first call.
Interviews by role type
Technical candidates should expect coding and, for engineers, system design. Anthropic's rule is no AI in take-homes or live interviews unless it says otherwise. One documented exception: a January 2026 Anthropic engineering post says candidates can use AI tools on its performance-engineering take-home.
Research scientist forms ask for a publications link. Applicants without publications are told to apply for the Research Engineer posting instead.
For non-technical roles, the careers page says interviews are conversational and focus on how you think through problems and what draws you to the work. If you need accommodations, the same page asks you to email [email protected].
What kinds of questions to prepare for
We don't publish "leaked" questions. Prepare for the categories Anthropic's own materials and the reports above point to: technical depth in your field, a detailed walk-through of past work, your motivation, and your values. interviewing.io lists a one-hour company-values session in the final round.
Can you use AI when applying to Anthropic?
Yes, within limits. Anthropic's guidance tells candidates to write the first draft themselves and then use Claude to refine it. AI isn't allowed in take-home assessments or live interviews unless Anthropic says so. The page was last updated July 10, 2025, and the application forms I opened ask you to confirm you've reviewed it.
| Stage | What Anthropic allows | What it doesn't allow | Source |
|---|---|---|---|
| Resume and application answers | Write your first draft, then refine it with Claude. Using Claude to describe real experience better, such as quantifying impact, is encouraged. | Using AI to invent experience or write answers about work you haven't done | Anthropic's AI guidance |
| Take-home assessments | AI only where Anthropic explicitly permits it | AI help by default, for example having it write code for a challenge | Anthropic's AI guidance |
| Interview preparation | Encouraged: research Anthropic, practice answers, prepare questions | Nothing listed | Anthropic's AI guidance |
| Live interviews | AI only if Anthropic says otherwise | AI assistance by default | Anthropic's AI guidance |
The rules were stricter before. In February 2025, Fortune reported that Anthropic's job postings carried an AI policy asking candidates not to use AI assistants during the application process. The current guidance sets stage-by-stage limits instead. Read the AI question on your own form, since that's the version that applies to you.
The page also sets out three principles. Use Claude as a collaborator in telling your real story. Use it to refine your ideas, not replace them. And be transparent about how you used it, the same way Anthropic is about its own use. One encouraged prep prompt it gives is: "What does Anthropic value in candidates based on their public information?"
Anthropic also says how it uses Claude in hiring: to create job descriptions, develop interview questions, draft candidate communications, analyze hiring metrics, transcribe interviews and identify candidates to source. It says it doesn't use candidate data to train Claude or let Claude make hiring decisions.
What this means if you use an AI job tool
- Use tools to find, filter and track Anthropic roles and similar roles elsewhere.
- Write the first draft of your resume changes and every written answer yourself. Refine afterward.
- Don't let any tool auto-submit an Anthropic application with AI-written answers.
- Don't use AI to add experience you don't have.
Anthropic salaries by role
Anthropic posts a pay range on its job ads. In the 10 postings below, US ranges ran from $220,000 to $850,000 a year and London ranges from £225,000 to £630,000.
| Role family | Posting | Location | Posted annual range | Checked |
|---|---|---|---|---|
| Research | Research Scientist, Interpretability | San Francisco | $350,000–$850,000 | Nov 13, 2026 |
| Research | Research Engineer / Scientist, Alignment | San Francisco | $350,000–$500,000 | Nov 13, 2026 |
| Research | Research Engineer, Pretraining | London | £260,000–£630,000 | Nov 13, 2026 |
| Engineering | Software Engineer, Claude Code | NYC, San Francisco, Seattle | $320,000–$560,000 | Nov 13, 2026 |
| Engineering | Staff Software Engineer, Infrastructure | London | £255,000–£325,000 | Nov 13, 2026 |
| Product | Product Manager, Revenue Platform | NYC, San Francisco, Seattle, remote-friendly | $305,000–$385,000 | Nov 13, 2026 |
| Policy | Public Policy, Strategic Advisory | San Francisco, Washington, DC | $265,000–$295,000 | Nov 13, 2026 |
| Policy | Product Policy Manager, User Well-being | San Francisco, NYC | $220,000–$255,000 | Nov 13, 2026 |
| Sales | Strategic Account Executive, Tech | San Francisco | $290,000–$380,000 (OTE) | Nov 13, 2026 |
| Applied AI | Applied AI Engineer | London | £225,000–£240,000 | Nov 13, 2026 |
London postings list pounds, and we haven't converted them, because exchange rates move and Anthropic doesn't publish a dollar equivalent for UK roles.
For software engineers specifically, the Claude Code posting lists $320,000–$560,000 in the US, and the London staff infrastructure role lists £255,000–£325,000.
Postings label the figure "Annual Salary." For sales roles, Anthropic says the range is on-target earnings, meaning base salary plus the commission or bonus target. Some postings add that total compensation also includes equity and benefits. Postings also say required experience depends on internal job level, which helps explain ranges as wide as the interpretability role's.
So don't read the top of a range as a likely offer. Ask your recruiter which level you're being considered for, and compare like with like: a sales OTE figure includes variable pay, while most engineering and research figures here are salary only. Pay data on employee-report sites can differ from these numbers. We use posted ranges because Anthropic publishes them and you can check them yourself.
The Fellows program, internships and other ways in
Anthropic's careers FAQ says it doesn't currently offer internships. The main alternative is the Anthropic Fellows Program, a funded research program with mentorship from Anthropic researchers.
The Fellows posting I checked describes four months of full-time work, with a possible extension, usually on an empirical project aimed at a public output such as a paper. It lists a weekly stipend of $3,850 (£2,310 or CA$4,300) and compute funding of about $15,000 a month. You need work authorization in the US, UK or Canada, and Anthropic says it can't sponsor visas for fellows. Applications run through Constellation, Anthropic's recruiting partner for the program, on a rolling basis.
You don't need a PhD, prior ML experience or published papers, according to Anthropic's Alignment Science blog. A full-time offer isn't guaranteed. The posting says "over 40% of fellows received a full-time offer" in previous cohorts.
The posting also says Anthropic's visa and office rules for full-time staff don't apply to the program, which changes the math for people who can't relocate to an office city.
The program suits engineers and researchers who want to move into AI safety research and can commit four months full time. It's a poor fit if you need visa sponsorship or a permanent role right away.
If you're an engineer without publications, the research engineer track is another route. The interpretability posting itself sends candidates without publications there.
How to write an application that stands out
Tailor your resume to one role
Pick one posting and read its "You may be a good fit if you" list. For each line that matches your background, find one piece of real evidence and put it where a reader sees it fast. Use numbers only when they're true and you can explain them. LiftmyCV's resume generator can help you adapt a draft you wrote to one job description. Edit what it gives you, line by line.
Illustrative example, not from a real candidate
Before: "Worked on evaluation tooling for the ML team."
After: "Built a Python evaluation harness that cut our ML team's regression-test run from a full day to under two hours."
For structure, see our machine learning engineer resume example or the product manager resume sample.
Recruiter's note from Marina Galkina
When I open a resume for a technical role, I start with the most recent job: the title, the stack and what the person actually built. Then I scan the bullets for scope and results. "Worked on the API" tells me very little. "Rebuilt the billing API and cut failed payments" tells me what you owned. If I can't see the match with the posting quickly, the resume has a problem.
Write the "Why Anthropic?" answer yourself
This is the field Anthropic flags as important. The interpretability form goes further and asks for "deeper and more specific engagement" with its research agenda, not general interest in AI.
Specific means you could only have written it yourself. Name a paper you read and the follow-up you'd want to test. Describe how you use Claude in your work and where it falls short. Or take a position on one of Anthropic's policy views and explain why. Praise for the mission alone tells the reader nothing new.
Recruiter's note from Marina Galkina
A strong “Why us?” answer could only have come from you. It names something specific, like a paper you read, a product you use every day or a decision the company made, and connects it to your own work. A generic answer praises the mission and could go to any company. I can usually tell which kind I’m reading in the first couple of sentences.
Show your work
The Alignment form asks for a link to the piece of work most relevant to the team. Research scientist forms ask for a link to your publications. Put GitHub repos, papers, technical writing or side projects in those fields and near the top of your resume.
Referrals and reaching out
interviewing.io reports that referred candidates may skip the coding challenge. Ask people who know your work, and give them a short note on the role and why you fit, so they don't have to write it for you. Don't chase recruiters with repeated messages, and don't ask contacts to pass along confidential details about the process.
Watch for scams too. Anthropic's job postings say its recruiters contact candidates only from @anthropic.com addresses and never ask for money or banking details before your first day.
Common mistakes to avoid
- Letting AI write your answers. Anthropic's guidance asks for your own first draft and rules out invented experience, and the form makes you confirm you've read it.
- Generic mission praise. The interpretability form explicitly asks for specific engagement, not general interest in AI.
- Applying to the wrong track. Research scientist forms send candidates without publications to research engineer roles.
- Skipping the office question. Anthropic's postings expect staff in an office at least 25% of the time, and the form asks.
- Overstating AI safety experience. Interviews go deep into your past work, and claims you can't back up will show.
- One resume for every role. A policy posting and a pretraining posting ask for very different evidence, so a single version will undersell you for at least one of them.
- Trusting unofficial recruiters. The same postings say real recruiters write only from @anthropic.com addresses and never ask for money.
Recruiter's note from Marina Galkina
The mistake I see most often from strong candidates is leaning on the names on their resume. They list well-known employers and assume the reader will fill in the rest. But I need to see what you personally did, at what scale, and why it matters for this role. Strong candidates often lose out to people who simply explain their work more clearly.
How LiftmyCV can help
Some parts of an Anthropic search suit a tool well. Use AI job matching to find roles at Anthropic and other AI labs that fit your experience. Keep every application, status and follow-up in the job application tracker. Use the resume generator to adapt a resume you wrote to one posting, then review the result yourself.
Matching is most useful beyond Anthropic. One company's openings come and go, and the same skills often fit research, policy or go-to-market roles at other AI labs. A wider list gives you more than one place to apply while you wait on a long process.
Sources and how this guide was researched
Who: Marina Galkina, Senior HR Manager, Lead Tech Recruiter, and Career Consultant at LiftmyCV. LiftmyCV is not affiliated with Anthropic.
How: I checked Anthropic's careers page and its guidance on candidates' AI usage on November 13, 2026. On the same day I checked the Fellows posting and 12 full-time job postings, 10 of them linked in the salary table. LiftmyCV's job data was pulled on November 14, 2026. It counts Anthropic's Greenhouse postings that are still live, without duplicates. Teams are grouped from job titles. Interview stages and timelines from interviewing.io and Exponent are attributed in the text. The LiftmyCV content team helped research and edit this guide. I checked every fact against its source.
Why: to help candidates prepare for Anthropic's process. LiftmyCV sells job-search tools, so this guide says where they help and where they shouldn't be used.
FAQ
How hard is it to get a job at Anthropic?
Anthropic doesn't publish an acceptance rate or applicant numbers, so nobody outside the company can give you a real figure. What it does publish is a written application with motivation questions, and the Claude Code application form says Anthropic values those answers highly.
Do I need a PhD to work at Anthropic?
No. Full-time postings ask for a bachelor's degree in a related field or equivalent experience. Anthropic's careers page says it cares about what you can do, not where you learned it. Research scientist roles do ask for publications, but research engineer roles don't.
Can I use Claude or ChatGPT to apply to Anthropic?
Partly. Anthropic's AI guidance lets you refine a draft you wrote yourself with Claude and encourages AI for interview prep. AI isn't allowed in take-homes or live interviews unless Anthropic says otherwise, and you mustn't use it to invent experience.
How long does Anthropic's hiring process take?
Anthropic doesn't publish a timeline. interviewing.io puts the process at about 3–4 weeks, while Exponent says most candidates report 1 to 4 months from application to offer. Ask your recruiter for your expected timeline on the first call.
Does Anthropic sponsor visas?
Yes, for many roles. Anthropic's careers FAQ says it sponsors visas and green cards, though some operations roles are harder to support, and postings say it can't sponsor every candidate. The Fellows Program is different: the Fellows posting says it can't sponsor visas for fellows.
Does Anthropic hire remote employees?
Rarely. Anthropic’s postings say it expects all staff in one of its offices at least 25% of the time, and some roles require more. A few postings are marked remote-friendly with travel, and the interpretability research role considers remote work case by case.
Read Next

How to Get Into Graphic Design and Land Your First Design Job
Choose a design lane, build a three-project portfolio with case studies, and apply while 9,064 graphic design roles are open.

How to Find a Job: The Complete Step-by-Step Guide
How to find a job step by step: where to search, how to get past the ATS, the best way to apply faster, and how to automate your job search with LiftmyCV.

How to Become a Copywriter, From First Sample to First Paid Job
Pick a copy lane, build four samples that prove it, and apply while 1,944 copywriter roles sit open. Pay data and timelines included.

