Software Engineer, Compute Infra HPC

Description

The company is hiring a Member of Technical Staff, Software Engineer for Compute Infra and HPC, to build the software that brings new frontier AI compute online and keeps it running reliably at global scale.

The work sits where distributed systems meet AI supercomputing. You will design control planes, APIs, workflow engines, state models and hardware-aware automation for some of the largest GPU systems in the world. The aim is to replace recurring manual operations with platform capabilities that shorten time to production, improve fleet availability and speed up researchers.

The team works on cluster bring-up and automation, fleet and node lifecycle and health, and automated operations at global scale. The systems turn newly delivered capacity into secure, observable, schedulable clusters and recover usable capacity when hardware or software fails.

The role is hands-on and based on the team's technical direction. The company works on its next generation of models with product teams that reach billions of users.

Responsibilities

• Design and build distributed services, control planes and platform APIs that cover the full lifecycle of AI compute clusters, from capacity intake and provisioning through certification, upgrades, maintenance, repair and decommissioning.
• Build foundational primitives that make new capacity easy to bring online, repeatable and secure by default. These include cluster bootstrap, Kubernetes control planes, networking, identity and access, image distribution, secrets, infrastructure as code and cluster metadata, across Azure and partner clouds.
• Own software that cuts cluster bring-up time by automating rack and node qualification, topology validation and scale testing, so new hardware moves to a healthy, schedulable fleet faster.
• Develop the node and fleet health platform. It combines telemetry, health signals, diagnostics and lifecycle state to detect, isolate and fix unhealthy hardware, hosts, networks and storage with little human involvement.
• Build policy-driven and agent-assisted automation for certification and maintenance, and for safe rollout and rollback of software and platform components.
• Use production data, benchmarks, workload profiles and fault injection to improve architecture, efficiency, failure detection, recovery time and the quality of health and readiness decisions.
• Turn incidents and repeated operational work into lasting software, clearer abstractions, automated guardrails and well-defined ownership, instead of one-off manual fixes.
• Work closely with researchers, hardware health, networking, storage, security and datacenter teams to shape the long-term architecture of the company's AI compute infrastructure.
• Lead technically through design reviews, architecture decisions, code quality, mentoring and delivery of complex cross-team initiatives.
• Live the company's culture and values.

Requirements

• Bachelor's degree in Computer Science, Computer Engineering, Electrical Engineering or a related technical field, plus 4+ years of software engineering experience building production distributed systems, infrastructure platforms or control-plane services, or equivalent experience.
• 4+ years designing and implementing scalable software for cloud, datacenter, cluster or fleet infrastructure in at least one general-purpose language such as Go, Rust, C++, C, Java or Python.
• Experience with Kubernetes or similar cluster orchestration systems, Linux, public cloud infrastructure, and infrastructure as code or declarative configuration.
• Proven ability to debug complex issues across services, control planes, operating systems, networks and hardware, and to turn what you find into sturdier software.
• Experience leading multi-team projects and explaining technical trade-offs clearly to both engineers and non-engineers.
• Preferred: Master's degree in Computer Science or a related technical field, plus 6+ years of software engineering experience building distributed systems or infrastructure platforms, or equivalent experience.
• Preferred: Experience running compute infrastructure at hyperscale, such as fleets of thousands of nodes, multiple clusters, several accelerator generations, or capacity spread across multiple cloud and datacenter providers.
• Preferred: Experience with cluster and node lifecycle systems, including bare-metal provisioning, cluster bring-up, health checking, certification, maintenance, repair automation, capacity management or decommissioning.
• Preferred: Deep knowledge of Kubernetes internals or related orchestration technology, such as controllers and operators, the scheduler, autoscaler, kubelet, Cluster API, container runtimes or workflow engines.
• Preferred: Experience building cloud foundations, such as virtual networking, CNI, private connectivity, DNS, workload identity, RBAC, secrets, image and artifact distribution, and Terraform or equivalent tools.
• Preferred: Experience designing health, diagnostics and remediation platforms using event-driven architectures, durable workflows, state machines, telemetry pipelines, automated decisions and safe recovery mechanisms.
• Preferred: Low-level systems experience with Linux kernels, device drivers, firmware, BMC or Redfish, secure boot or attestation, virtualization or hardware health agents.
• Preferred: Experience with GPU or accelerator systems and high-performance interconnects such as InfiniBand, RoCE or RDMA, NVLink, NCCL or collective communication, plus high-performance storage and data movement.
• Preferred: Experience owning production reliability and incident response while removing recurring toil through software and automation.
• Preferred: Commitment to clean, testable, maintainable, secure and well-documented code, with sound judgment on correctness, scalability, operability and long-term platform evolution.
• Preferred: Ability to align senior stakeholders, mentor engineers and drive multi-quarter technical initiatives from architecture through production rollout.
• Preferred: Enthusiasm for learning fast-moving infrastructure technology and for making very large compute systems more reliable, efficient, understandable and easy to use.

Conditions

Level: Software Engineering IC5.

UK base pay range: £93,500.00 to £161,800.00 per year.

Expected to work from a designated Microsoft office at least four days a week if you live within 50 miles (U.S.) or 25 miles (non-U.S., depending on country) of that office. This depends on local law and may differ by jurisdiction.

The position stays open for at least 5 days, and applications are taken until it is filled.

Benefits

• Some roles may include benefits and other compensation. Further benefits and pay details are available through the company's listing.