Agent Harness Engineer (On-site)

Viktor; München; Vor Ort; EUR 120.000 - 160.000; Vollzeit

Zusammenfassung

Viktor is hiring a role focused on building and evolving an AI agent runtime that plans, acts, and verifies its output. You’ll integrate across many tools, expand skills, and push reliability with evals while shipping features to users through Slack and Teams.

This in-office role requires fluency with agentic workflows, testing, and scalable design. You’ll work directly with founders in a fast-moving, high-trust team environment.

Qualifikationen

• You’ve built or significantly modified an AI agent harness.
• Ability to ensure loop closure with tests, linters and verifications.
• Experience creating custom skills, commands, CLIs, or MCP servers.
• Comfort with end-to-end agent runtime, deployment and evaluation.

Aufgaben

• Build the agent runtime loop of planning, acting, and verifying work.
• Make the agent fluent with tools and recall across integrations.
• Grow agent capabilities with memory and autonomous tasks.
• Improve reliability via evals and regression checks.
• Ship product features to users through Slack and Teams.
• Collaborate in a small team to cover a wide surface.

Kenntnisse

AI agent development

Agent integration experience

Automation testing

Systems thinking

End-to-end automation

Tools

TypeScript/React

Python

CLI tooling

Jobbeschreibung

The Short Version

You’re the person who makes Viktor do more things, for more customers, more reliably. You’ve built agents before (runtime, tools, memory, evals) and have opinions about what makes them work. You ship the day you write the code and reach for AI-assisted development by default. If you’ve never built an agent, this isn’t the role.

The Hard Part

Viktor runs a persistent Linux sandbox for every workspace and writes its own Python to drive 3,000+ integrations, from Salesforce to Stripe to Shopify. That’s 1.5M+ tool calls a day against real customer data, and the curve is steep.

We’re betting one general agent, fluent across everything a company runs on, beats a stack of narrow tools. An agent that impresses in a demo and one that 25,000+ teams trust with their real work are different machines, and the distance between them is the whole job.

What You’ll Actually Do

• Build the agent runtime: the loop where it plans, acts, and verifies its own work instead of declaring success it didn’t earn.

• Make the agent fluent with tools: composition across many integrations, recall, and the skill files that keep the right functions at hand.

• Grow what the agent can do: memory it can trust, skills it writes for itself, work it runs on its own.

• Make it measurably better: evals, catching regressions, and turning every failure into a fix.

• Ship product: features that reach users through Slack and Teams within hours.

• Whatever needs building. Small team, large surface.

The Bar

You ship to production every day, and the changelog has your name on it. When Viktor breaks at 3am, you can fix it because you understand the system end to end. This role doesn’t work without agentic coding fluency: if AI-assisted development isn’t already how you work, you’ll be behind on day one.

Who You Are

• You’ve personally built or significantly modified an AI agent harness, and can describe the trade-offs you made.

• You’ve made an agent reliably close the loop: tests, linters, typecheckers, browser verification, whatever stops it hallucinating success.

• You’ve built custom skills, commands, CLIs, or MCP servers to make your own agentic coding faster.

• Agentic coding is your daily workflow, and you’ve stress-tested models and harnesses across the frontier.

• Systems thinking. You make hard technical trade-offs and design for how things break at scale.

• Speed, with the bar up. You ship today, not Thursday.

• Range. Agent runtime to React component to deployment script in an afternoon.

• Genuine interest in how AI works. You’ve read the papers and have opinions about evals.

• Onsite. This is an in-office role, not remote.

Why This Role Is Different

• No layers. You work directly with both founders, and decisions get made in the room, not in a Linear ticket.

• The agent is the product. Every reliability win shows up in retention the next week.

• Volume that forces you to be good. 1.5M+ tool calls a day means the lazy answer breaks in production immediately.

Even Better If

• Previous AI engineering experience: agents, harnesses, eval infrastructure.

• Previous founder or early-stage builder.

• Open source contributions to the AI tooling we live in.

Tech

TypeScript/React on the frontend. Python on the backend and agents. Modal for infrastructure. You don’t need all of it coming in, but you need to learn fast.

How we work

Small team, high trust, low process. Decisions are made by owners, not committees. You will ship your first week. You will talk to users your first day.

We don’t do alignment meetings or stakeholder syncs. We build things, see if they work, and iterate.

Why Viktor

We’re one of the fastest-growing companies in the world. The product works. The market is pulling.

This is a rare window: everyone here owns something real. Not a task. A surface of the company that customers depend on.

That doesn’t last forever. Right now, it’s still true.

Compensation

Top-of-the-market salary and the kind of ownership that only exists at this stage.

The best work happens when you’re in the room. Munich, New York, Dublin, and Warsaw. Remote for some roles.