Anthropic Residency 2027: Paid AI Research Placements | Salary $35,0000 to $ 500,000 per year equivalent

This is not a student programme. The Anthropic Residency is a paid AI research programme for people who have already led big technical work. They now want to move into AI safety research.

Placements last 6 weeks or 3 months or 6 months. Pay is set at $350000 to $500000 a year equivalent. You do not need a PhD or machine learning experience or published papers.

Applications for this round close on Sunday 22 November 2026 at 11:59 pm Pacific Time. Offers are expected in January 2027.

Last Information Checked: 4 November 2026

At a Glance

Item	Details
Programme	Anthropic Residency
Run by	Anthropic
Type	Paid fixed term research programme
Focus	AI safety and alignment and AI security and ML systems
Placement length	6 weeks or 3 months or 6 months
Pay	$350000 to $500000 a year equivalent
Location	San Francisco preferred (remote is possible for shorter placements)
Who it is for	Professors in quantitative fields and technical founders and staff level engineers
PhD or ML experience	Not required
Start date	Flexible (the programme aims to arrange it)
Deadline	22 November 2026 at 11:59 pm PT (07:59 UTC on 23 November)
Offers	Expected in January 2027

About the Anthropic Residency

Anthropic gives a simple reason for this round. Machine learning is a young field. Many senior researchers started as physicists or mathematicians or neuroscientists or software engineers.

As models improve the key skill is to frame a research plan and judge what is worth doing. Senior people in quantitative fields often do this well already. The Residency helps them learn enough about frontier AI to use those skills in a new field.

This round builds on the Anthropic Fellows Program. Past fellows worked on subliminal learning and the Petri auditing tool and persona vectors. The Residency adapts that model for established technical leaders.

How the Residency Works

• A matched project: Residents work on problems outlined by Anthropic researchers. The problems are matched to your interests and background.
• A mentor: Each Resident is paired with an Anthropic mentor and gets regular one on one guidance.
• Time to learn: You are not expected to arrive as an ML expert. The programme includes self study time and talks and guidance on modern ML and AI safety research.
• Compute: Residents work with frontier models and have a dedicated compute budget.
• Flexible timing: You say which length you prefer when you apply. The programme aims to arrange flexible start dates.
• Location: Anthropic prefers that Residents work from its San Francisco offices. Remote work is possible especially for shorter placements. Most full time research roles at Anthropic need a move to San Francisco.
• The goal: By the end every Resident should be ready to continue AI research full time at Anthropic or elsewhere.

Also See: ETH Zurich Summer Research Fellowship 2027

What the Pay Means in Real Terms

The official pay is a yearly equivalent. A short placement pays only for the time you stay. Here is simple maths if pay is spread evenly over time.

• 6 weeks: about 40000 to 58000 dollars
• 3 months: about 87500 to 125000 dollars
• 6 months: about 175000 to 250000 dollars

These are my estimates and not official figures. The official text gives only the yearly rate.

Research Areas

AI safety and alignment

• Scalable oversight: keeping very capable models helpful and honest
• Adversarial robustness and AI control: keeping advanced systems safe in unfamiliar or hostile settings
• Model organisms: building controlled examples of misalignment to study how failures arise
• Model internals and interpretability: understanding how large language models work inside
• Safeguards research: defences against attacks and safety evaluations and tools that catch risks before release
• AI welfare: studying possible AI welfare issues and building evaluations

AI security and frontier red teaming

• Stress testing frontier models to learn what they can do in areas such as cybersecurity and robotics and cryptography

ML systems and reinforcement learning

• Improving how frontier models are trained. This covers architecture and optimisers and training data and hardware and reinforcement learning environments.

Potential Mentors

• Sara Price leads the Alignment Training team
• Ethan Perez leads the Alignment Science team
• Evan Hubinger leads the Alignment Stress Testing team
• Samuel Marks leads the Scalable Oversight team
• Elana Simon studies the interpretability of language models and biological models
• Belinda Li studies how language models represent the world and users and themselves
• Jack Lindsey leads the Model Psychology team
• Nicholas Carlini studies adversarial machine learning and AI security
• Jascha Sohl-Dickstein works where safety meets pretraining

Their backgrounds include data science and mathematics and physics and neuroscience. This shows the programme values people from other fields.

Residency Benefits

• A scoped project matched to your background
• Direct mentorship from Anthropic researchers
• Access to Claude models and GPU compute
• A dedicated compute budget for your project
• Self study time and research talks
• Pay at a $350000 to $500000 a year equivalent
• A link to the wider AI safety and security research community
• Flexible placement length and start dates
• Support to continue AI research full time after the programme

Who Can Apply

Anthropic is looking for people who:

• Have led ambitious work in a technical field. These are professors in a quantitative discipline or technical founders or staff level engineers.
• Care about making AI safe and beneficial for society
• Want to move into empirical AI research full time
• Do well in fast paced and team based settings
• Can put ideas into practice quickly and communicate clearly

You do not need a PhD or machine learning experience or published papers. Anthropic also encourages applications from groups that are underrepresented in tech. Applicants may also be considered for the Anthropic Fellows Program.

Hiring Process

1. Initial application
2. Recruiter screen
3. Technical assessment
4. Research discussion
5. Reference check

Anthropic expects to make offers in January 2027.

How To Apply Step By Step

1. Read the official Anthropic Residency announcement.
2. Choose the posting that fits you. Apply through the US posting or the UK posting.
3. Decide your preferred placement length. The application asks for it.
4. Fill in the application. The announcement does not list the required documents so check the posting.
5. Submit before 22 November 2026 at 11:59 pm Pacific Time.

Key Dates

• 22 November 2026 at 11:59 pm PT: Application deadline (07:59 UTC on 23 November and 12:59 pm Pakistan time on 23 November)
• January 2027: Offers expected
• Start dates: flexible and not fixed