Senior DevOps Engineer

Pravar AI builds governed agent infrastructure for wealth management firms. Our customers are financial services firms, RIAs and asset managers who will demand varying levels of control of their data - On-Prem, Client VPC, Airgapped, or Multi-tenant. These require enterprise grade security, they can see, permission, and audit every action it takes. That requirement shapes how we deploy. Each customer comes with a variety of deployment needs, cost optimization and security.

You would own the machinery that makes this repeatable. The job is to take a stack with many moving parts, including agent runtimes, policy and redaction layers, evaluation pipelines, and vector stores, and make it install cleanly in places we don't control. You would also keep it running once it's there. We are a small, fast team, so you will have real influence over how we build, and you will ship things in your first month.

• Design and run the pipelines, infrastructure code, and release process that take Pravar AI from commit to customer environment.
• Write Terraform and Helm packaging that behaves the same way on AWS, Azure, GCP, and bare-metal Kubernetes.
• Build deployment paths that hold up under customer change-control and approval gates.
• Treat the air-gapped install as a design constraint from the start, not a special case added later.
• Manage observability and reliability across those environments: metrics, logs, traces, alerting, and runbooks.
• Own the SLOs we commit to, and lead the incident reviews that follow when we miss one.
• Own backup, failover, and disaster recovery, and test them on a schedule rather than trusting the diagram.
• Build vulnerability scanning, dependency checks, and image validation into the pipeline.
• Manage secrets and certificates, and support customer-managed encryption keys.
• Help collect the evidence our SOC 2 and ISO/IEC 42001 work requires, and make that collection routine.
• Build the tooling behind our delegated support model: just-in-time access, session recording, and least-privilege roles.
• Work at the cutting edge of product availability and enterprise requirements, with room to try new things.

• Four to six years in DevOps, platform engineering, or SRE, with at least two years running production systems that other people depended on.
• Real experience deploying in both worlds. You have shipped to public cloud and to customer-controlled or on-premises infrastructure, and you know how different the two feel when something breaks at 2 a.m.
• Strong Kubernetes and Helm skills, including ingress, autoscaling, network policy, and upgrades without downtime.
• Infrastructure as code in practice. Terraform or something equivalent, written for reuse and not just for a single environment.
• CI/CD fluency in GitHub Actions, GitLab CI, Jenkins, or ArgoCD, along with sound release habits: versioning, promotion, rollback.
• Working knowledge of observability (Prometheus, Grafana, OpenSearch, or comparable) and a feel for what is worth alerting on.
• Solid Linux, networking, and scripting. You can debug DNS, TLS, and IAM problems without a long detour, and you write Bash or Python when a task calls for it.
• A security-minded habit of mind. Least privilege, secrets handling, and encryption in transit and at rest are things you do by default.
• Comfort with ambiguity and pace. You make a reasonable decision with incomplete information, write down why, and revisit it when you learn more.

• Experience in regulated settings such as banking, fintech, or healthcare.
• AWS/ GCP DevOps certification.
• Experience of running on-prem infrastructure.