Software Engineering Manager, Network Security

Anthropic•San Francisco, CA

1d•$405,000 - $485,000•Hybrid

About The Position

Anthropic's mission is to create reliable, interpretable, and steerable AI systems. The Network Security team's remit is to secure Anthropic's networks across our cloud providers and our bare-metal AI data centers. We own egress and ingress controls, network segmentation, private connectivity between environments, and encryption in transit. We design, build, and maintain these controls, and we write the software that deploys, enforces, and monitors them. These controls protect our model weights, customer data, and internal assets. We are hiring an Engineering Manager to lead this team. Anthropic is adding data center sites and cloud regions quickly, and the team will grow with that work. In this role you will set the team's charter, practices, and priorities. You will partner with our infrastructure networking, cluster, and data center teams so security is built into network designs from the start. The goal is secure-by-default networking that lets research and product teams use compute in any environment without slowing them down.

Requirements

8+ years of front line management experience
Low ego, high empathy; has built diverse teams and has former reports who would work for them again
Experience managing infrastructure or security engineering teams, including hiring, developing, and managing senior and staff engineers
Has grown engineers into staff or lead roles, and has made the call to move someone out when coaching didn't work
Kept a team intact and delivering through reorgs, charter changes, or leadership turnover
Experience owning a security or infrastructure domain end to end: setting multi-year strategy, delivering year-long cross-organizational work with little direction, and answering to executives for its risk and tradeoffs
Scaled a team and its systems together through hypergrowth, where last quarter's architecture was already too small
Hands-on network security engineering in cloud and/or on-prem environments: VPC design, interconnects, private service connectivity, cloud firewall policy, egress control, physical segmentation, and out-of-band/management-plane security
Threat modeling and risk assessment for networked systems, with the judgment to tell which controls reduce risk and which just look good on a diagram
Secured Kubernetes networking at scale, or worked in high-assurance environments where controls had to be evidenced, not just deployed

Nice To Haves

Built or rebuilt a hiring loop and held the bar while adding headcount quickly
Exposure to large-scale distributed training and what it demands of a network
Solved end-to-end networking problems directly with internal users, cloud account teams, and hardware or colo vendors
Has taken infrastructure from zero to one and then to maturity: built the first version where nothing existed, then run and hardened it into a system other teams depend on at production-network reliability
Pragmatic and steady: cares about unblocking research and product as much as a clean architecture, and brings patience and judgment to opinionated senior engineers
Clear and persuasive communicator in writing and in person
Familiar with AI security and safety concepts

Responsibilities

Own the network security roadmap: threat model from the management plane through Kubernetes and cloud interconnects, decide what to build, and deliver it
Be able to reason about the design and implementation of default-deny ingress and egress, segmentation between research, training, and production, private connectivity, and mTLS/service identity across every environment
Ship these as secure-by-default primitives so new clusters, VPCs, and data center links inherit the right boundaries without security on the critical path
Join networking, cluster, and data center teams at design time so topology, routing, and firewall decisions carry our invariants from day one
Build visibility into traffic and boundary drift, with detection for exfiltration paths and automated remediation where safe
Own the network security evidence for our security goals and compliance: the invariants hold, and you can prove it
Hire, grow, and retain senior security software engineers; own their career development, and set the practices the team runs on as it scales