Getviktor

Viktor is the AI coworker. It lives in Slack and Microsoft Teams, connects to thousands of tools, and does real work for real companies: finance, marketing, ops, engineering. We're building the product that replaces half the SaaS stack with a single teammate.
The team is small. The scope is not.
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 you have opinions about what makes them work. You ship to production the day you wrote the code, you're fluent in agentic coding workflows, and you reach for AI-assisted development by default because that's how you're fastest.
If you've never built an agent, this isn't the role. We're hiring people who already have the muscle and want to build the agent everyone else will try to copy.
What's Actually Going On Here
Someone in Slack asks Viktor to reconcile their Stripe payouts against their books. Viktor does it, live, in under a minute. The customer tells their network. Two more teams sign up that week. That's the loop. Your job is to make it happen more often, across more surfaces, more reliably.
Viktor handles 600K+ tool calls a day and the volume curve is steep. We connect to thousands of tools - Salesforce, Notion, Stripe, QuickBooks, HubSpot, Shopify, whatever customers ask for next. Each new integration unlocks a new shape of customer. Each reliability win compounds.
What You'll Actually Do

• Build the agent runtime. Sandboxed execution, tool orchestration, memory, the closed loop where the agent checks its own work. The part of Viktor that turns a sentence in Slack into a real action against a real API - and verifies it landed.
• Ship integrations. OAuth, webhooks, schema mapping, error handling. Each new integration unlocks a new shape of customer. The bar is that it works on the customer's actual data on day one, not on a fixture.
• Run the infrastructure. Container orchestration, autoscaling, cost per task. Viktor stays fast and cheap because you're paying attention to it.
• Ship product. Features that go directly to users through Slack and Teams. You'll see people using what you built within hours.
• Whatever needs building. Small team, large surface. You'll work on things that aren't in this document.

The Bar
You ship to production every day, and the changelog has your name on it. A feature you built is the reason a customer closed. When Viktor breaks at 3am, you can fix it because you understand the system end to end. The founders are writing less code because you've taken over surfaces they used to own.
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.
Day-1 Reality
Week 1: ship something to production. Take ownership of a live surface.
Who You Are

• You've personally built or significantly modified an AI agent harness. You can describe the trade-offs you made and why.
• You've made an agent reliably close the loop - tests, linters, typecheckers, browser verification, whatever it took to get the agent to check its own work instead of hallucinating success.
• You've built custom skills, commands, CLIs, or MCP servers to make your own agentic coding faster. You don't just use the tools shipped to you.
• Agentic coding is your daily workflow. Agents, parallelized workflows, whatever makes you fastest. You have informed opinions about which models, harnesses, and tools to reach for and when - and you've stress-tested across the frontier, not just settled on one.
• Systems thinking. You make hard technical trade-offs and own them. You understand how things break at scale and design for it. Good judgment about what's worth doing well and what's worth doing fast.
• Speed, with the bar up. You ship today, not Thursday. No enterprise engineering. No three-week review cycles.
• Range. You can go from the agent runtime to a React component to a deployment script in an afternoon. Viktor's surface area is too big for specialists.
• Genuine interest in how AI actually works. You've read papers. You've read other people's prompts. You have opinions about how evals should work.
• Remote. Work from wherever you're sharpest. We default to async, ship in public channels, and bring the team together in person a few times a year.

Why This Role Is Different

• No layers. You work directly with both founders on the surface that matters most. Decisions get made in the room, not in a Jira ticket.
• The agent is the product. You're not bolting AI onto something - you're building the thing customers came for. Every reliability win, every new tool, every memory improvement shows up in retention the next week.
• Volume that forces you to be good. 600K+ tool calls a day means the lazy answer breaks in production. You learn what actually scales because the system tells you immediately.
• The feedback loop is hours, not quarters. Ship in the morning, customer uses it in the afternoon, you see what worked by end of day. If that pace is what you've been missing, this is the room.

Even Better If

• Previous AI engineering experience: agents, harnesses, deep model understanding, eval infrastructure.
• Previous founder or early-stage builder. You've worn every hat and liked it.
• 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 to know all of it coming in. 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.
Everyone here owns something real. Not a task. A surface of the company that customers depend on. When it breaks, you fix it. When it wins, everyone knows whose work it was.
We use Viktor to build Viktor. You'll see what you're working on in action every day.
Why Viktor
This is a rare window. The product works. The market is pulling. The team is small enough that what you do next week will be live in production next week. That doesn't last forever. Right now, it's still true.
Compensation
Competitive salary and the kind of ownership that only exists at this stage.
We're in Munich, New York, and Warsaw. Onsite preferred. The best work happens when you're in the room.