Staff Engineer, On-Prem Security Platform

Qevlar AI; United States; Hybrid; USD 150,000 - 210,000; Full time

Job summary

Qevlar AI is seeking an advanced engineer to own on-premise deployment for customers who require data to stay inside their perimeter. You will design reference architecture, ensure air-gapped installation, and replace cloud services with self-hosted equivalents across Postgres, object storage, and secret management.

You will optimize the inference stack for GPU throughput, build a robust test harness, and enforce security through image signing and SBOMs, with a focus on Kubernetes and OpenShift

Qualifications

• Shipped software that others install and operate in restricted environments.
• Deep Helm expertise with subcharts and OCI registries.
• Experience in air-gapped environments with offline bundles and proxies.
• Replaced managed cloud services with self-hosted equivalents.
• Practical software supply chain security including image signing and SBOM.
• Strong IaC and CI experience (Terraform, Ansible).
• Hands-on experience serving models on GPU with VRAM sizing.
• Strong Go and/or Python coding skills.
• Ability to exercise architectural authority and say no.
• Fluent in French and English.

Responsibilities

• Define on-prem reference architecture and list non-portable dependencies.
• Ship the bundle with Helm charts and air-gapped tarball for Kubernetes/OpenShift.
• Replace cloud with self-hosted Postgres, S3-compatible storage, and secret management.
• Size inference layer with VRAM and concurrency sizing; manage model weights distribution.
• Build test harness for environment- and release-testing across clusters.
• Ensure images are signed, SBOMs generated, and base images minimal.
• Design upgrade paths with feature flags and health checks.
• Maintain a simple, auditable self-hosted deployment without new managed dependencies.

Skills

Helm expertise

Kubernetes experience

Go programming

Python programming

CI/CD pipelines

Infrastructure as code

GPU model serving

Security & SBOM/CI

Air-gapped deployment

Go/Python coding

Tools

OpenShift

OCI registries

Vault / External Secrets

Terraform

Ansible

NVIDIA GPU Operator

Container Toolkit

cosign

MinIO

Postgres operator

Job description

Qevlar AI is seeking an advanced engineer to own on-premise deployment for customers who require data to stay inside their perimeter. You will design reference architecture, ensure air-gapped installation, and replace cloud services with self-hosted equivalents across Postgres, object storage, and secret management.

You will optimize the inference stack for GPU throughput, build a robust test harness, and enforce security through image signing and SBOMs, with a focus on Kubernetes and OpenShift