Elan is a platform for building, deploying and scaling AI agents. A developer writes an agent, and the platform provides everything else needed to run it as a product. You would be one of the first two engineers on the team.
As DevOps and Deploy Engineer you own the path from a developer's deploy command to a running, routable, policy-compliant workload. You build the deploy tooling as a product, and you operate the infrastructure under it.
What you will own
Deploy tooling. A deploy service and CLI that turn a declarative manifest into a running workload, with status, logs, rollback and clean teardown.
GitOps. A Git-driven configuration repository, reconciled to the clusters with Argo CD or Config Sync.
Kubernetes and GKE. GKE Autopilot across separate production and non-production projects, with namespace isolation per workload, Workload Identity, autoscaling and scale-to-zero.
Agent sandboxing. Every customer agent runs as untrusted code, isolated under gVisor (GKE Sandbox), with resource limits and no route to other tenants.
Networking. Ingress and routing, domains, DNS, TLS certificate automation and egress control.
Configuration and secrets. Managed across environments and tenants.
Infrastructure as code. Terraform for shared and per-workload infrastructure on GCP, with AWS as a second cloud.
CI/CD and workflow automation. Container image builds for Python, Node.js and Go workloads, an artifact registry, vulnerability scanning, image signing, and canary and batched rollout. Repetitive engineering and operations work automated with GitHub Actions and scripts, so nothing is done by hand twice.
Package releases. Publishing and versioning the platform's SDKs and CLI on PyPI through automated release pipelines, extending to npm as SDKs for other languages are added.
Platform security. Admission control with OPA Gatekeeper, Binary Authorization, NetworkPolicy and default-deny egress, built to SOC 2 standard.
Reliability and cloud spend. Monitoring, alerting, backups and restore tests. Optimising cloud spend through right-sizing, autoscaling, scale-to-zero and committed-use discounts, with cost reported per workload and per tenant.
How you will work
You report to the head of product and technology.
You work beside a Platform Engineer who owns the agent runtime and the shared services. The two of you alternate on-call weeks, and you lead infrastructure incidents.
You are based in Noida with a hybrid working arrangement. Compensation is 40 to 90 LPA.
We hire for four things: ownership, curiosity, developer obsession and delivering results.
Minimum requirements
8 to 15 years in infrastructure, DevOps or platform engineering, with at least 3 running Kubernetes in production. GKE preferred, EKS acceptable.
Builds product as well as pipelines: has shipped a service, API or CLI that other developers depend on, in Python or Go.
Terraform at module-authoring level, including state management across environments.
GitOps in production with Argo CD, Flux or Config Sync.
Has designed a CI/CD pipeline from scratch, including image builds for more than one language runtime and promotion between environments.
Networking in depth: Kubernetes ingress or Gateway API, DNS, TLS certificate automation and VPC design.
Kubernetes security hands-on: admission control with Gatekeeper or Kyverno, NetworkPolicy, workload identity, and image signing.
Has run untrusted or multi-tenant workloads in sandboxed containers on Kubernetes. gVisor (GKE Sandbox) strongly preferred.
Has published and maintained packages on PyPI, with automated, versioned releases.
Has cut a cloud bill measurably, and can explain where the savings came from.
Automates by default: writes tooling and workflows in Python or Bash, and on GitHub Actions or a similar CI system.
GCP fundamentals: IAM, Cloud SQL, Pub/Sub and Secret Manager.
Has carried a pager, led incident response and written postmortems.
Good to have
Has taken infrastructure through a SOC 2 or ISO 27001 audit.
Experience building an internal developer platform or a deploy product.
KEDA, or egress proxies such as Envoy.
Experience as an early engineer at a startup.
AWS.
