Product & AI Quality Engineer

Evos • New York, NY • Technology, Information and Internet • 2d ago

Save this role or tell us whether you want more jobs like it.

Own the release gate for an autonomous engineering org — you decide what “ready” means, and nothing ships without the evidence.

Evos reached a $1M run rate four weeks after launching, selling to small and mid-sized businesses in physical industries, among the hardest markets to sell software into. Roughly 3% of startups reach $1M inside a year at all. Our initial traction puts us among the fastest-growing companies in history, and we have no intention of slowing down.

Our thesis is simple. Software is becoming the execution layer, not a system of record, and not a tool that waits to be prompted. Legacy sectors account for more than half of global GDP, and the software sold into them has always had the same ceiling: someone still has to manage it, instruct it or prompt it. Evos is changing that.

We codify the expertise of a company’s best operators, connect to their tech stack, and deploy Evos operators that execute the same way those people do. Evos takes over full job roles in complex businesses and executes, so they don’t have to. In these industries, the knowledge that keeps a company running walks out the door when its best people leave. Evos makes it permanent.

Our mission is to change how work gets done in the industries that move the world. If you want a say in what that future looks like, Evos is the right place to be.

Why work at Evos

Evos is the execution layer for critical industries, and you decide what’s ready to reach them. This isn’t software that assists or drafts; it does the actual work, live inside the businesses that move the world. You’re the person who calls whether a release is fit to go into production. Your judgement is what stands between the work and the companies that depend on it.
You have the final say on the product before it ships. Our engineering runs autonomously and code moves faster than any human team could review it, which means your role is to own what “ready” means.

The role

Our engineering function is autonomous, shipping fixes and features at very high volume. Your job is to make sure what it produces is actually ready before it reaches the product. That takes meticulous attention to detail, a clear sense of what the product should look like in production, and a real understanding of our customers, so you can tell the difference between what technically passes and what’s genuinely right for them. You decide what “ready” means, you catch what isn’t, and you’re the one who says not yet.

What you’ll own

The release gate. You define what “working as designed” means for every flow a customer touches, and you hold releases to it. Nothing ships on inference, including yours: you show the failing check when you block, and save the evidence when you pass. When something ships broken, you’re the first to know and the one who explains why.
Making your judgement scale. Your eye becomes automation — Playwright E2E coverage for the flows that matter, and checks that run without you. You’ll write and maintain these in TypeScript, and build internal tools where the existing ones fall short.
Root-cause on customer bugs. You trace a reported bug to a reproduction and a cause, and write it up so an engineer fixes it in one pass.
Front-to-back verification. What works on the front end has to actually work underneath it. You cross-check the two, find where they diverge, and isolate the specific failure — then identify it, resolve it, and prove it’s fixed rather than assuming it is.

The technical environment

An autonomous engineering org. Code ships faster than people do. Your QA has to keep pace with a system that never stops moving.
An AI-executed QA harness. Much of the release gate is already run by AI agents, with saved evidence per check. You’ll direct it, extend it, and decide what it checks.
A config-driven platform. Behavior loads from configuration rather than living in separate codebases, so a change in one place can surface anywhere. Knowing where to look is half the job.
Core stack: TypeScript, React, Convex, Sentry, Clerk Auth, Inngest, Playwright, Vercel.

Who you are

You’ve QA’d fast-moving, AI-heavy products. Systems that ship multiple times a day and don’t behave identically twice — where “it worked once” isn’t a pass. You know how to define “passing” for something that changes underneath you, and hold that standard without slowing the ship rate.
You have taste. The layout a few pixels out, the flow that works but feels wrong, the state nobody handled — you catch what technically passes but isn’t right, and you know which of it blocks a release and which goes on the list.
You’re meticulous, and you don’t cut corners. You won’t take the shortcut that patches an issue when there’s a longer route that resolves it for good. You’ll nitpick every detail until it’s right — a perfectionist, in the way the job needs.
You can build what you need. Solid TypeScript, comfortable writing Playwright E2E tests and building a small internal tool when the existing ones fall short. You don’t need an engineer to make your checks run.
You drive QA with AI. You use AI agents for the repetitive parts, check their work, and reject it when the evidence isn’t there.
High agency. You own the quality outcome, not a ticket queue — you decide what needs checking, and you’re comfortable being the person who holds up a release.

Nice to have

Prior exposure to logistics, manufacturing, or another physical-industry enterprise.
Design sensibility — you can tell a UI is wrong before you can say why.