Full-Time
Staff Software Security Engineer
Posted on 9/3/2026
Anthropic
5,001-10,000 employees
Develops reliable, interpretable AI systems
Compensation Overview
£255k - £325k/yr
London, UK
Hybrid
At least 25% of working time is expected in an office.
Bachelor's
 Cybersecurity (1) |
|---|
Get referred to Anthropic
See people who can refer or advise you
- Have 8-15+ years of software engineering experience implementing and maintaining critical systems at scale.
- Have strong software engineering skills in Python or at least one systems language such as Go, Rust, C, or C++.
- Have experience implementing and operating critical systems at scale using DevOps and cloud automation practices such as infrastructure as code.
- Have hands-on network security engineering experience in cloud or on-premises environments, including areas such as VPC design, cloud firewall policy, egress control, private service connectivity, or network segmentation.
- Have working knowledge of Kubernetes security and networking, including network policy, service identity, and container hardening.
- Have experience with threat modeling and risk assessment for networked and multi-cloud systems.
- Have a track record of driving engineering excellence through high standards, constructive code reviews, and mentorship.
- Be able to communicate clearly, translate technical concepts across organizational levels, and bring clarity and ownership to ambiguous technical problems.
- Support fast-paced startup engineering teams.
- Have a bachelor's degree or an equivalent combination of education, training, and/or experience.
- Have a field of study relevant to the role, as demonstrated through coursework, training, or professional experience.
- Build security for large-scale artificial intelligence clusters across multiple clouds and bare-metal data centers, including identity and access management, network segmentation, and encryption controls.
- Design and implement network security controls including default-deny ingress and egress, segmentation between research, training, and production environments, private connectivity across clouds and data centers, cloud firewall policy, and mutual transport layer security with service identity.
- Ship security controls as secure defaults using infrastructure as code and GitOps workflows so new clusters, virtual private clouds, and data center links inherit the appropriate boundaries.
- Build visibility into network traffic and boundary drift, with detection for exfiltration paths and automated remediation where safe.
- Design secure-by-design development workflows and continuous integration/continuous delivery pipelines across services, with expertise in Kubernetes security, container orchestration, and identity management.
- Threat model and assess risk for complex multi-cloud and networked environments, and join infrastructure networking, cluster, and data center teams at design time so topology, routing, and firewall decisions incorporate security requirements from the outset.
- Mentor engineers and contribute to hiring and growth of the Security team.
- Experience securing Kubernetes networking at scale, including container network interfaces, network policy, service mesh, or mutual transport layer security rollouts, or experience in high-assurance environments where controls had to be evidenced rather than merely deployed.
- Experience with interconnects and private connectivity between cloud providers and physical data centers, including management-plane and out-of-band network security.
- Experience building network traffic visibility or exfiltration detection systems.
- Exposure to large-scale distributed training and its network requirements.
- Experience leading cross-functional security initiatives and navigating complex organizational dynamics.
- Experience managing infrastructure through automated configuration and policy enforcement.
- Experience hardening containerized applications and enforcing security policies.
- Experience attracting talent and supporting diverse, inclusive teams.
Anthropic focuses on AI research to build reliable, interpretable, and steerable AI systems. Its main product, Claude, is an AI assistant designed to handle tasks at any scale for clients across industries, delivered through deployment and licensing along with specialized AI R&D services. Claude works by combining natural language processing, human feedback, reinforcement learning, and interpretability techniques to produce a capable, controllable AI assistant that can assist with a wide range of tasks. The company differentiates itself from competitors by prioritizing safety, transparency, and controllability—emphasizing reliability, interpretability of model behavior, and user-controlled steerability in its AI systems. Anthropic’s goal is to make AI systems that people can trust and efficiently use to improve operations and decision-making across sectors.
Company Size
5,001-10,000
Company Stage
Debt Financing
Total Funding
$182.8B
Headquarters
San Francisco, California
Founded
2021
Get referred to Anthropic
See people who can refer or advise you
Simplify's Take
What believers are saying
- Project Glasswing found over 10,000 critical vulnerabilities by June 2, 2026.
- Anthropic disclosed annualized revenue above $65 billion in July 2026, signaling explosive demand.
- September 1, 2026 pricing cuts for cache reads boost agentic API adoption and retention.
What critics are saying
- Anthropic’s $1.5 billion copyright settlement, approved July 20, 2026, invites more suits.
- Late-September 2026 IPO pressure exposes weak multiples if growth decelerates after listing.
- Heavy compute commitments and chip-lease debt create existential financing risk if demand softens.
What makes Anthropic unique
- Claude Security and Project Glasswing anchor Anthropic’s enterprise security moat in 2026.
- Anthropic pairs frontier-model capability with explicit safety branding, unlike OpenAI’s consumer-first posture.
- Multi-cloud distribution across AWS, Google, Microsoft, Lambda, and Nscale reduces single-vendor dependence.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Flexible Work Hours
Paid Vacation
Parental Leave
Hybrid Work Options
Company Equity
Growth & Insights and Company News
6 month growth
↑ 7%1 year growth
↑ 4%2 year growth
↑ 2%Broadcom could capture a $40 billion opportunity from Anthropic's growing compute needs, according to Macquarie analyst Arthur Lai. This comes as concerns mount over Google developing more chips internally, potentially threatening Broadcom's custom silicon business. The stock has fallen approximately 24% from its all-time high. However, Lai suggests many concerns may already be priced in, creating an attractive entry point. In April, Anthropic partnered with Google and Broadcom to secure next-generation TPU capacity for training its Claude AI models. Broadcom's AI semiconductor revenue surged 221% year-over-year to $16.7 billion in fiscal Q3 2026. Management projects AI semiconductor revenue could reach $115 billion in fiscal 2027 and potentially $230 billion in fiscal 2028. The Anthropic partnership could provide crucial revenue visibility whilst strengthening Broadcom's position in custom AI silicon and networking.
Black Duck has joined Anthropic's Project Glasswing, an industry initiative aimed at securing critical software infrastructure using advanced AI for defensive cybersecurity. The application security company will apply Mythos, Anthropic's AI system, across its full security portfolio. This will combine AI-accelerated vulnerability discovery with remediation workflows, risk-based prioritisation, and compliance-driven governance. "AI is transforming the economics and speed of vulnerability discovery and exploit development," said Dipto Chakravarty, Black Duck's Chief Product & Technology Officer. He explained that pairing Mythos with Black Duck's existing capabilities will enable faster risk reduction whilst maintaining the transparency and auditability required by enterprise security teams. Black Duck specialises in application security, combining deterministic analysis with AI reasoning to identify and fix security issues in code written by developers, generated by AI, or assembled from open source.
Goldman Sachs and Morgan Stanley have asked major credit rating agencies to grant investment-grade status to OpenAI and Anthropic upon going public, despite neither company turning a profit, the Financial Times reported. OpenAI posted a $20.9 billion operating loss on $13.1 billion revenue in 2025. Anthropic doesn't expect to break even until 2028, with OpenAI targeting 2030. The investment-grade designation would allow pension funds and insurers to buy their bonds. It would also terminate Nvidia's guarantee of up to $105 billion in lease obligations for OpenAI's Ohio campus. Rating analysts currently describe both labs as speculative-grade and loss-making. When SpaceX received investment-grade ratings after its June IPO, its bonds traded near junk pricing within days. Anthropic could list in late September, whilst OpenAI targets 2027.
Interactive Brokers held $182.4 billion in uninvested client cash at the end of June, up 27% year over year, and this figure grew to $185.6 billion by August. The automated global broker earns interest on this cash by investing it in short-term US government securities whilst paying clients a rate half a percentage point below the federal funds rate. Net interest income rose 23% year over year to $1.06 billion in the second quarter, representing more than half of total net revenues of $1.9 billion. The growth came from larger balances rather than margins, which actually narrowed to 1.93% from 2.07%. Anthropic's potential IPO, rumoured to arrive soon with a possible $2 trillion valuation, could provide clients with an opportunity to deploy some of this cash.
Anthropic has reportedly secured a $35 billion cloud deal with Lambda for 350 MW of capacity at Hut 8's Beacon Point campus in Texas, marking its ninth major compute corridor. The arrangement highlights Nvidia's dual role as both GPU supplier and data centre landlord, allowing it to extract value at multiple levels. The deal supports Anthropic's $65 billion annualised revenue run rate but deepens its reliance on Nvidia's ecosystem. Anthropic has diversified across nine corridors, including commitments to AWS (5GW), Google/Broadcom (5GW), Microsoft/Nvidia ($30 billion), Fluidstack ($50 billion), Nscale ($45 billion), Volta ($10 billion), AMD ($5 billion), and SpaceX (300MW). This infrastructure strategy reflects a shift where GPU suppliers increasingly control both hardware and physical environments, positioning themselves as compute landlords.