Product Engineer

About the role

ABOUT OPSLEVEL

OpsLevel is the agentic delivery engine designed to help dev teams build, ship, and maintain software better than ever. The way software gets developed and delivered is changing rapidly, and OpsLevel provides the context and guardrails that your human & agent developers both need to navigate their world. Entire engineering teams at great companies like Okta, Keller Williams, Nerdwallet, and Zapier run on OpsLevel.

We're a fully-remote team of team members across North America. OpsLevelers live our company values on a daily basis and when we’re not heads-down on a project or collaborating in real-time over Zoom, we're competing in trivia over lunch or sharing the latest fur baby photo in #pets-to-make-you-live-longer.

Our culture and values are very important and ground us in how we interact as one team:

Growth Mindset - Get comfortable with being uncomfortable

Focus & Efficiency - Keep the main thing the main thing, and get it done

Empathy - Think of others; think of your users

Intellectual Honesty - The best solution is more important than anyone's ego

Simplicity - Don't overthink it

ABOUT THE ROLE

This is a product engineering role. You'll own problems from end to end: understanding what customers are stuck on, shaping the solution with product and design, building it, and shipping it. Ownership doesn't stop at the merge. You're the one watching how it gets used and deciding what the next version looks like.

You'll also work with agents every day. Everyone here routes work through coding agents, and we're building the internal pipeline that makes that actually work: issues structured so an agent can pick them up, context files that give agents what they need, and guardrails that catch them when they get it wrong. Yes, we dogfood! We're doing to ourselves exactly what our customers are trying to do to their own engineering orgs, and a lot of our best product ideas come out of it.

OUR STACK

We like boring technology because it's well tested, simple, and well supported. It's also what agents are best at. Decades of public Rails code in the training data means agents need less correcting here than on newer frameworks. Our current stack:

- Ruby on Rails (our main application's backend)

- Vue.js (our app's UI)

- Golang (for operational tools and our open source offerings)

- GraphQL (how the UI and the backend talk to each other; customers also use it as part of our platform)

- MySQL (gotta store those bits somewhere)

- Redis (some bits gotta be stored faster)

- Terraform (config as code, because any other way is painful)

- Kubernetes (we use AWS EKS)

- AWS / GCP

- Docker

- DataDog

On the agent side we're less settled, on purpose. Claude Code, Cursor, Codex and more are in daily use, and we expect the specifics to change a few times a year. Whatever the tools, we treat agent workflows as part of the product, not a side experiment.