# Principal Site Reliability Engineer, Infrastructure Architecture- AI Infrastructure Operations — Nscale

- Location: Houston, TX
- Posted pay: $150k–$250k/yr
- Posted: Apr 3, 2026
- Status: open (checked Nov 21, 2026)
- Apply (employer's site): https://job-boards.eu.greenhouse.io/nscaleoperationsukltd/jobs/4774936101
- Employer record: https://visajobs.io/company/nscale-operations-us.md

## Sponsorship history
- Tier: Some history
- Occupation (SOC): Software Developers (15-1252)
- Employer's certified LCAs for this occupation in TX, FY2024–FY2026: 1
- …in the whole US: 6
- With this exact title: 0
- Median DOL prevailing wage, level I / II: $84k / $106k

## Description
About Nscale
Nscale is the GPU cloud engineered for AI. We provide cost-effective, high-performance infrastructure for AI start-ups and large enterprise customers. Nscale enables AI-focused companies to achieve superior results by reducing the complexity of AI development. Our GPU cloud bolsters technical capabilities and directly supports strategic business outcomes, including cost management, rapid innovation, and environmental responsibility.
We thrive on a culture of relentless innovation, ownership, and accountability, where every team member takes pride in their work and drives it with excellence and urgency. As an Nscaler, you’ll build trust through openness and transparency, where everyone is inspired to do their best work. If you join our team, you’ll be contributing to building the technology that powers the future.
About The Role
Nscale’s AI Infrastructure Operations team keeps one of the most demanding AI platforms in the industry running. At our scale, the design decisions we make this year set the failure modes we live with for the next five.
We’re looking for a Principal Site Reliability Engineer, Infrastructure Architecture to be the architectural authority for how our platform is designed, connected, and recovered.
This is a design-first role. You will own the reference architecture for Nscale’s GPU, network, and control-plane platforms; decide where failure domains and trust boundaries sit; and make sure every major system we build has a known blast radius, a tested recovery path, and a clear owner. You'll work alongside the teams doing bring-ups, upgrades, and migrations, and you'll join major incidents when the cause turns out to be in the design. You'll be a peer to our existing architecture lead, not a backup; the two of you share ownership of both the design and the operating model, so neither one is a single point of failure.
What You'll Be Doing
Owning the end-to-end reference architecture for Nscale’s AI and HPC infrastructure: compute, network fabric, storage, control plane, and the core services underneath them
Designing multi-site and multi-region topologies, including failure domains, tenant isolation, and secure service enclaves for sovereign and hyperscale customers
Separating control plane from data plane by design, so that a management-layer outage never takes customer workloads down with it
Defining the target-state architecture and the multi-year roadmap to reach it, including what we retire, consolidate, or rebuild
Leading architecture reviews across Engineering, Network Operations, and Fleet Operations; writing and approving the RFCs and architecture decision records (ADRs) that other teams build against
Setting the architectural standards every new system must meet before it ships: SLOs, observability, capacity models, dependency maps, and documented failure modes
Running pre-mortems and red-team reviews on major designs to find single points of failure and hidden coupling before production does
Joining major incidents when the root cause is architectural, and follow the fix through to the design
Co-owning the operating model with the SRE team: operational readiness reviews (ORR), incident severity and post-incident review (PIR) standards, and production access and break-glass policy
Making sure the teams running each system can deploy, upgrade, and recover it without needing you in the room
Planning and reviewing site bring-ups, fleet upgrades, and migrations of live customer workloads
Running DR tests and game days so we know recovery works
Using incident and on-call data to decide what to redesign next
Mentoring senior engineers in systems design, so that architectural knowledge and decision-making never sit with a single person
About You
10+ years building and running large-scale production infrastructure, including on-call for systems you designed, and several years as a staff, principal, or architect-level engineer
A track record of architectures you designed that are running in production today; you can walk us through the tradeoffs, what broke, and what you would change
You've led major incidents and run large migrations or site bring-ups
Working knowledge of change management, staged rollouts, SLO-based alerting, and post-incident reviews
Deep expertise in distributed systems design: consistency models, consensus and replication, partition handling, and graceful degradation
Strong networking fundamentals at the design level, including L3 fabrics, BGP, EVPN/VXLAN, segmentation, and multi-site connectivity
Experience designing secure multi-tenant platforms: identity, PKI and secrets management, network isolation, and zero-trust boundaries
Clear technical writing; your design docs, RFCs, and ADRs hold up under scrutiny, and engineers can build directly from them
Proven ability to drive architectural decisions across teams without direct authority, including saying no to designs that will not hold up
Strong software engineering skills; you can prototype a design, review production code, and spot when an implementation has drifted from the intended architecture
Sound judgement when reliability, velocity, and cost pull in different directions, and the ability to explain the tradeoff to engineers and executives alike
Nice to Have
Architecting AI or HPC platforms at scale, including GPU clusters, InfiniBand or RoCE fabrics, and workload schedulers such as Slurm or Kubernetes-based GPU scheduling
Designing sovereign, regulated, or air-gapped cloud environments
Bare-metal and hybrid cloud architecture, including provisioning pipelines, a CMDB/DCIM as the source of truth, and hardware lifecycle automation
Observability architecture for high-cardinality, high-throughput environments
Kubernetes platform architecture across many clusters and sites
Structured design analysis methods such as failure mode and effects analysis (FMEA), threat modelling, or C4 diagrams
A history of step-change reliability or scalability gains that came from a redesign rather t

> History, not a promise: past H-1B filings do not mean an employer will sponsor you or a given role. Data: U.S. DOL OFLC LCA disclosure (FY2024–FY2026 for role counts, certified LCAs) and the USCIS H-1B Employer Data Hub. Not legal advice. Methodology: https://visajobs.io/methodology
