Founding Infrastructure Engineer

Modern Relay; United States; On-site; USD 140,000 - 200,000; Full time

Job summary

Modern Relay is building Omnigraph's control plane and deployment platform. We seek an infrastructure engineer who has built control planes for data platforms, with Terraform, Kubernetes, and GitOps as your first languages.

You will own the control plane and deployments across cloud, on‑prem, and hybrid environments, delivering a sovereign, one‑click experience for customers to run models anywhere. Join a founder-led team shaping the future of the platform and enabling customers with a flexible,

Qualifications

• Hands-on Kubernetes control-plane work.
• Declarative IaC with Terraform as a first language.
• Built data-platform infrastructure from design to implementation.
• GitOps as a default — infrastructure changes are code reviews, not tickets.
• Security depth — supply-chain / zero-trust thinking, identity and SSO.
• Track record designing reliable, correct systems with clear failure semantics.

Responsibilities

• Own the control plane, including cluster creation and visibility.
• Deployments across hosted cloud, customer VPC, on‑prem, and hybrid architectures.
• Drive security and identity integration (SSO) and policy engineering.
• Develop GitOps workflows and IaC standards for environments.
• Collaborate with founders to define the infrastructure strategy.

Skills

Kubernetes control-plane

Terraform / IaC

GitOps practices

Security depth

Reliability & fault tolerance

Agent-native development

Ownership mindset

Tools

Kubernetes

Terraform

Job description

Design, build, and operate the components that provide storage, control-plane, and policy services across cloud, hybrid, and on-prem environments.

You will own the layer that turns Omnigraph from an engine into a platform: the control plane and every way it gets deployed. We're hiring an infrastructure engineer who has built control planes for data platforms: from design to implementation. Terraform, Kubernetes, and GitOps are your first languages.

Sovereignty is our core differentiator: customers run Omnigraph in our cloud, their cloud, or their data center, with full control of which model sees which data. Your job is to make that promise real, repeatable, and one click away.

We're building

A lakehouse-native graph engine with git-style workflows. Branch, commit, and merge typed graph data like code. Multi-modal, self-hosted, open source. Optimized for on-prem and hybrid context graphs, multi-agent research, and enterprise knowledge systems.

We believe graphs are not just a database choice. They are the most efficient way to create and distribute knowledge — across people and agents. As fans of Deutsch's The Beginning of Infinity, we think we're at the start of an explosion of explanatory knowledge, and we want to lead the way.

Our mantra is SOVEREIGNTY. The engine is live and clients are deploying. What's missing is the platform layer that makes every deployment effortless. We invite you to build it with us.

Our Style

We've optimized the whole company to run on agents: Hermes, Claude Code, Devin — Slack wired for agent-driven development. You won't hand-crank infrastructure.

You’ll build it to run on rails: declarative, GitOps-driven, agent-operable, cluster lifecycles as code.

What you’ll own

Four systems to build and run:

Control plane — the cluster control plane for Omnigraph: one‑click cluster creation, prewarmed cluster pools, branch/query/graph visibility, multi‑tenancy, usage and cost observability. Take it from working prototype to sellable product.

Deployments — every shape a customer needs: hosted cloud, customer VPC, on‑prem, hybrid. Cluster cloning for client sandboxes and pilots; the Docker/CI rails underneath; object‑storage‑first architecture (S3 and friends) end to end.

Security & identity — supply‑chain integrity and zero‑trust posture; identity, SSO, and access policy; secrets and credential flows. Drive the design and development of policy‑engine integration (Cedar) into cloud‑native authorization environments. Enterprise buyers ask hard infosec questions — you'll have the answers and the architecture behind them.

DevOps / DevEx — GitOps workflows, IaC standards, CI/CD, environments. The internal platform that lets a small team (and its agents) ship like a big one.

Must have

Hands‑on Kubernetes control‑plane work — not just consuming clusters, building the machinery that manages them

Terraform / IaC as a first language — you think in declarative infrastructure

Built data‑platform infrastructure from design to implementation — data stacks, data engineering context, and the ops reality behind them

GitOps as a default — infrastructure changes are code reviews, not tickets

Security depth — supply‑chain / zero‑trust thinking, identity and SSO, a real sense for cyber

Track record designing systems emphasizing reliability, correctness, and clear failure semantics

Agent‑native — you live in Claude Code / Codex / Open Code and orchestrate agents to scale your work

High agency, high ownership — you don't wait, you ship

Nice to have

Object‑storage internals (S3 semantics, conditional writes, S3‑compatible backends) and columnar / lakehouse formats (Arrow, Parquet, Iceberg, Delta, Lance)

Rust, or willingness to work at the engine boundary

Enterprise delivery experience: procurement, compliance, on‑prem installs behind strict perimeters

Time in OSS infra / dev‑tools communities

Who you'll work with

Founder team from day one. Direct access to the founders and the engineering team for architecture, escalation, and roadmap. You set the infrastructure strategy; we give you the access and the budget to run it.

Work

Equity: meaningful — this is one of the crucial early roles

Remote: remote‑first, with monthly offsites