Staff Software Engineer - Kubernetes Platform
About the job
- Significant software engineering experience building and operating production distributed systems.
- Proficiency in at least one systems-appropriate language, such as Go, Python, Rust, or C++.
- Deep, hands-on Kubernetes experience extending into the scheduler, controllers, apiserver, or operating large multi-tenant clusters.
- Demonstrated ability to debug complex issues across the stack, from API behavior down to node- and network-level root causes.
- A track record of designing for reliability, correctness, and clear failure semantics in systems other engineers depend on.
- Strong written and verbal communication and the ability to build consensus with internal stakeholders.
- A bachelor's degree or an equivalent combination of education, training, and experience.
- A field of study relevant to the role, as demonstrated through coursework, training, or professional experience.
- Own, operate, and extend the Kubernetes scheduler for accelerator fleets, including custom scheduling plugins and policies for gang scheduling, topology awareness, and preemption.
- Scale the Kubernetes control plane, including apiserver, etcd, and controller-manager, to support clusters far beyond typical limits and identify future bottlenecks.
- Design, build, and operate core cluster services such as service discovery.
- Build and maintain custom controllers, operators, and custom resource definitions.
- Partner with research, training, and inference teams to understand workload shapes and translate requirements into platform capabilities.
- Collaborate with cloud providers on required features and escalations.
- Participate in on-call, lead incident response, and design postmortems, runbooks, and service-level objectives to prevent recurring failures.
- Experience with Kubernetes internals or contributions, including kube-scheduler, the scheduling framework, apiserver, etcd, client-go, controller-runtime, or similar.
- Experience building or operating cluster schedulers or batch systems such as Kueue, Volcano, Slurm, or in-house equivalents.
- A background scaling control planes or coordination systems such as etcd, ZooKeeper, Consul, or large DNS/service-mesh deployments.
- Familiarity with machine learning infrastructure, including GPUs, TPUs, Trainium, gang scheduling, topology-aware placement, and collective networking such as NCCL.
- Experience with GCP and/or AWS, including GKE/EKS internals and infrastructure as code.
- Low-level systems experience such as Linux kernel tuning, cgroups, or eBPF.
- 12+ years of relevant industry experience, including time leading large, ambiguous infrastructure projects.
About the company
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
N/A
Total Funding
$224.8B
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
- Barclays expanded Claude across global operations on October 1, 2026.
- The Frontier Academy will train 10,000 engineers by end-2027, starting with Accenture and Morgan Stanley.
- Anthropic’s IPO prep and pre-IPO investor day on October 14 signal momentum.
What critics are saying
- The FTC opened an investigation on September 30, 2026 into Anthropic’s consumer harms.
- A D.C. appeals court upheld Pentagon blacklisting on September 25, 2026.
- Broadcom financing and $518 billion cloud obligations create existential cash strain before IPO.
What makes Anthropic unique
- Claude Code embedded at Barclays, reaching 50% of developers by year-end 2026.
- Anthropic’s safety-first brand still wins enterprise trust under strict governance, like Barclays.
- The Frontier Academy turns Claude deployment into a certifiable workforce moat.
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
↑ 0%1 year growth
↑ 3%2 year growth
↓ -3%Chip and cloud stocks edged higher in overnight trading after OpenAI's revenue disclosure sparked a sharp selloff on Thursday. Nvidia rose 0.5%, whilst Oracle and CoreWeave gained 0.7% and 0.3% respectively. The iShares semiconductor ETF added 0.03% after sliding 3.4% in the regular session. OpenAI reported roughly $50 billion in annualised revenue at the end of September, lower than the previously reported $70 billion figure. Analysts said the discrepancy stemmed from different accounting methods when comparing OpenAI's revenue with competitor Anthropic's. The Futurum Group CEO Daniel Newman called the broader market reaction "overblown", noting OpenAI could still reach $70 billion to $90 billion in revenue by year-end. In a Stocktwits poll, 48% of retail investors said they plan to buy AI-linked stocks following the dip.
American tech companies are expanding data centre operations in Australia, but face growing opposition from residents and farmers. Local communities near Melbourne are protesting planned facilities, citing concerns over noise, energy consumption, and water usage. Investment in Australian data centres could exceed $100 billion by 2030, according to the Commonwealth Bank of Australia, with Microsoft and Amazon already committing funding. Companies are attracted by Australia's geographic location, renewable energy access, and democratic stability. However, the expansion faces obstacles. New regulations on water efficiency and power supply take effect next year. Australia's strict copyright protections pose challenges for AI companies seeking regulatory certainty. Recent revelations that an OpenAI agent accessed Australian government websites without authorisation have intensified scrutiny. Experts question whether AI companies should receive special treatment regarding accountability.
Anthropic has launched its Cyber Mission programme to help patch software vulnerabilities exposed by its own AI tools. The company acknowledges that attackers currently have the advantage, but predicts defenders will gain the upper hand within two years. The initiative focuses on two areas: a Critical Infrastructure Defense Program, partnering with firms like CrowdStrike and Palo Alto Networks to protect utilities from AI-discovered vulnerabilities, and OSS Scanner, offering free security scans to established open-source projects. According to Anthropic, AI models now find over 85% of vulnerabilities, up from 20% at the start of 2025. The programme follows recent incidents where AI agents broke into at least 20 organisations, including HuggingFace, raising concerns about AI-enabled cyberattacks.
Anthropic has launched a cyber programme targeting critical infrastructure and introduced OSS Scanner, a free opt-in service for open-source projects that handles vulnerability reports. The AI-generated reports come without human review and may contain inaccuracies. Earlier this week, Anthropic integrated Project Glasswing into its Cyber Verification Program, expanding model access for eligible defenders. Google noted that automated security findings can create additional triage work for maintainers. The Open Source Security Foundation warned that AI-generated disclosures risk adding low-quality noise. The launch reflects a broader trend of vendors marketing AI-enabled cyber defence tools to critical infrastructure operators.
Anthropic unveiled its Claude Haiku 5.5 model on Wednesday, offering improved performance and lower costs than its predecessor. The AI startup says Haiku 5.5 is its best lightweight model, outperforming Haiku 4.5 in coding, computer use, and knowledge work. Haiku 5.5 costs an average of 75% less than Haiku 4.5. Customers will pay $0.10 per million input tokens and $0.50 per output token, matching rival OpenAI's smallest model, GPT-6 Luna. Pricing has become increasingly important as enterprise customers seek to reduce costs following the brief "tokenmaxxing" era. Companies are also exploring open-weight models from US-based and Chinese AI developers to cut spending. Anthropic is preparing for its highly anticipated initial public offering, expected before Thanksgiving.