Staff Network Development Engineer

About Nscale

Nscale builds GPU cloud infrastructure for AI. Our engineering teams develop the infrastructure and software used to deploy and operate large-scale AI systems.

About the team

Network Development and Applications builds the software that helps Networking deliver consistent, repeatable deployments. We own the applications for provisioning, configuration generation and validation, and work with fabric teams, Network NPI, Deployment Engineering and SRE to make them reliable and supportable.

About the role

As a Staff Engineer, you will lead a substantial domain within Network Development and Applications, such as provisioning, configuration tooling or validation. You will turn requirements from several teams into a coherent technical design and deliver it through supported releases. Your scope will typically span two to four collaborating teams or a defined application domain, with hands-on responsibility for the difficult parts of implementation.

What you will do

• Own the technical design and delivery plan for a substantial application domain. Break ambiguous requirements into practical milestones and document important decisions.
• Build reusable Python servicesand libraries, Ansible automation workflows,
• and integrations for provisioning, configuration generation, validation or service applications, working within the agreed portfolio architecture.
• Work with fabric engineers to define contracts for intent, configuration content and test criteria. Make errors diagnosable and preserve traceability from input to result.
• Lead integration with Network Source of Truth, ZTP, deployment tooling and Network Control Plane applications as appropriate. Manage compatibility, schema changes and staged migrations with the receiving teams.
• Develop test and release approaches covering realistic device behaviour, partial failures and recovery. Use labs or simulation with the teams that own them.
• Take proven tools through documentation, packaging, adoption and support readiness. Agree maintenance responsibilities before a workflow depends on them.
• Lead complex application troubleshooting, remove recurring causes of failure, review code and designs, and mentor engineers across the domain.

Candidate requirements and expected outcomes

What you will bring

• Experience leading the design and delivery of a substantial production application or network automation domain across collaborating teams.
• Strong Python engineering and practical experience with APIs, structured data, Git reviews, automated testing and CI/CD. Able to design maintainable interfaces and diagnose integration failures.
• Experience applying structured software development practices, including iterative planning, backlog prioritisation, incremental delivery and regular feedback, using workflows such as Kanban.
• A record of turning recurring requirements into reusable capabilities, including upgrades, documentation and support after release.
• Ability to lead delivery without line-management authority, make trade-offs clear and raise the quality of other engineers’ work.

Experience for your hiring profile

Network Development

Deep experience in at least one of network provisioning, configuration automation or validation. Comfortable with Linux, IP networking, device APIs and the operational risks of applying automation across a fleet. Experience with tools such as Ansible, Jinja2 and NetBox or equivalent.

Network Applications

Deep experience building production backend services or internal engineering applications. Comfortable with API design,SQL and relational databases such as PostgreSQL, MariaDB or MySQL, including schema design and migrations, asynchronous work and containerised delivery. Able to work with network specialists to model device workflows and test their failure modes.

Useful additional experience

Our current stack includes Python, Ansible, Jinja2 and NetBox, with FastAPI, NiceGUI, MariaDB, NGINX, ISC Kea and Docker Compose in ZTP tooling. Relevant experience with these or comparable tools is useful. AI/HPC networking, multi-vendor device integration, lab simulation and platform handovers are also valuable. You do not need expertise in every listed technology.

What success looks like

• A domain design and delivery plan agreed with its users and integration partners.
• A supported capability adopted by dependent teams, with reliable releases and clear troubleshooting guidance.
• Fewer recurring failures or manual steps in the domain, and engineers able to maintain it beyond its original author.