How much money Anthropic pays software engineers in London
Key Points
- Anthropic’s London software engineering roles pay from £325,000 a year.
- Three Staff-level infrastructure roles reach £485,000 a year.
- Five other listings cap pay at £390,000 a year.
- Staff must spend at least 25% of their time in the office.
- Anthropic sponsors visas but warns candidates about recruiter scams.
Anthropic is offering London software engineers between £325,000 and £485,000 a year, according to eight job listings recently added to its internal jobs board.
All eight roles start at the same £325,000 floor, and three Staff-level infrastructure posts go up to £485,000 a year. The other five listings stop at £390,000 a year.
At the top of the range, the pay works out to about £40,416.67 a month before tax. At the floor, it comes to about £27,083.33 a month.
The roles and the pay
These three listings carry the highest ceiling, and all of them deal with the hardware and clusters behind Anthropic’s compute fleet:
- Staff Software Engineer, Node Infra: £325,000 to £485,000
- Staff Infrastructure Engineer, Cluster Infrastructure: £325,000 to £485,000
- Staff Software Engineer, Kubernetes Platform: £325,000 to £485,000
These five listings are capped at £390,000 a year:
- Software Engineer, Observability & Profiling
- Staff Software Engineer, AI Reliability Engineering
- Staff Software Engineer, Continuous Integration
- Staff Software Engineer, Inference
- Staff Software Engineer, Infrastructure (Distributed Systems)
The Node Infra team brings in accelerator capacity from all the major cloud providers and from datacentres built for Anthropic. It also builds automation that finds and repairs faulty GPU, TPU and Trainium nodes.
The Kubernetes Platform team owns the scheduler and control plane that run the company’s training and serving clusters. It also builds the core services every workload in the fleet depends on.
“We are operating at a scale where the defaults stop working,” Anthropic said in its Kubernetes Platform listing.
The Inference team runs the systems that serve Claude to millions of users. These include request routing, autoscaling and the pipelines that release new models.
The AI Reliability Engineering team works with other teams across the company to keep Claude available. Its remit covers every step from the developer SDK through the network and serving layers to the accelerators.
Experience and education
The Kubernetes Platform listing asks for 12 or more years of relevant industry experience as a preferred qualification. It also asks for time spent leading large infrastructure projects.
Most of the other listings prefer candidates with at least 10 years of experience. Node Infra sets its preferred bar at seven years, and that must include time as a technical lead.
Anthropic asks for a bachelor’s degree, or an equivalent mix of education, training and work experience. It said it matches years of experience to internal job levels rather than setting a fixed minimum.
Anthropic encourages people to apply even when they do not meet every listed qualification.
Office time, visas, and benefits
Staff must work from one of Anthropic’s offices at least 25% of the time, and some roles may require more.
Anthropic said it sponsors visas but cannot do so for every role or candidate. It said it keeps an immigration lawyer on retainer to help people who receive an offer.
The listed benefits include optional equity donation matching, generous holiday and parental leave, and flexible working hours.
Applicants must answer a “Why Anthropic?” question, and the company said strong answers often run to between 200 and 400 words. Candidates must also confirm they have read Anthropic’s guidelines on using AI during the application process.
Anthropic warned that its recruiters only contact candidates from @anthropic.com email addresses. The firm said legitimate recruiters never ask for money, fees or banking details before a new hire’s first day.
