Senior SRE: GPU AI Platform Reliability & SLIs

Mirantis; United States; Hybrid; USD 140,000 - 230,000; Full time

Benefits offered by this job

Competitive compensation package

Strong benefits plan

Professional development and training

Conferences and hackathons

Job summary

Mirantis is seeking a Senior SRE to define reliable SLIs/SLOs for a GPU-accelerated AI platform and own the end-to-end reliability contract between the platform and its operators. The role spans Kubernetes, bare-metal, and NVIDIA infrastructure, building an API to surface SLIs for Platform Administrators and downstream systems.

The candidate should craft alerting strategies, manage error budgets, and collaborate across infra, storage, and networking to ensure actionable signals and low noise.

Qualifications

• 5+ years in SRE, platform reliability, or related role.
• Strong software engineering skills (Go or Python) with production APIs/services experience.
• Experience defining SLIs/SLOs and error budgets for real systems.
• Hands-on observability tooling experience (metrics, logs, tracing).
• Solid understanding of Kubernetes and its signals.

Responsibilities

• Define SLIs and SLOs across the platform using available signals.
• Design and build the API exposing SLIs and reliability state to admins and downstream systems.
• Establish alerting and error-budget practices to maximize signal-to-noise.
• Collaborate with infra, storage, and networking teams to instrument signals.
• Diagnose reliability and performance issues across the observability stack.

Skills

Go

Python

APIs

Kubernetes

Observability

Prometheus

Grafana

Tools

Prometheus

VictoriaMetrics

OpenTelemetry

Grafana

Job description

Mirantis is seeking a Senior SRE to define reliable SLIs/SLOs for a GPU-accelerated AI platform and own the end-to-end reliability contract between the platform and its operators. The role spans Kubernetes, bare-metal, and NVIDIA infrastructure, building an API to surface SLIs for Platform Administrators and downstream systems.

The candidate should craft alerting strategies, manage error budgets, and collaborate across infra, storage, and networking to ensure actionable signals and low noise.