Careers

Help us make customer support run itself

Productlane is the customer support platform for software companies, built on Linear. Five people in Munich, profitable, hiring two.

The bet we are making

In most software companies, support headcount grows with the customer count. We think that stops being true this decade. A company with 100,000 customers will answer all of them without a support team.

Productlane is what takes its place. The agent answers the customer directly. When it finds a bug, it files a Linear issue and follows up once the fix ships.

What is left after the agent is usually a real product problem. Some of it is relationship work that belongs with sales. The rest is software work that belongs with engineering.

We also think support is the software category that grows fastest from here. It is the one place you can hand an agent thousands of real conversations and count how many it got right.

The number to move

The agent closes 35% of conversations today. Both of these roles exist to get that to 80%.

Where a person steps in

When the agent cannot finish a conversation, it hands over to a person with the full history. Deciding when to hand over is the hard part.

Closing the loop with Linear

A customer message turns into a Linear issue. When the fix ships, the customer hears about it without anyone remembering to tell them.

The companies whose support already runs on Productlane

Clerk logo
Bitcoin.com logo
Flightcontrol logo
Ghost logo
Inngest logo
Apostrophe logo
Linearity logo
Speakeasy logo
Novu logo

Where we are

Our agent handles about 24,000 customer conversations a month and closes 35% of them without a human. We want that number at 80%. Most of what we build this year is aimed at it.

We bootstrapped to hundreds of users before raising $1M from the founders of Linearity, JOKR and MVST. We have been profitable since January 2026, and revenue is growing fast.

Most of that growth is usage-based, because we charge per resolved ticket. So when the agent gets better, we see it in the same month's revenue.

Because we are profitable, the roadmap gets decided by the five of us and the customers who write in.

Once the agent handles 80%, support stops being a team you hire for. We would like every company on Linear to run that way, and eventually every software company.

"Really cool see the tools other folks build on top of the @linear API. @raphael_167 has been one of the early users on Linear and quit his job to focus on this idea to make user insights better and more transparent."
Karri Saarinen
Karri Saarinen · Cofounder, Linear · on X

How we work

  • No PM layer. Engineers talk to customers and decide what to build.
  • Everyone owns a number. You'll know which one in your first week.
  • We write decisions down in Linear, so you can find out why something happened a year later.

Practical details

  • Hybrid, from our office in Munich.
  • Salary plus equity, for both roles.
  • 30 days of vacation, plus German public holidays.

Who you would work with

Five of us today. These two roles make it seven.

  • Raphael Fleckenstein
    Raphael FleckensteinFounder, CEO
  • Danylo Shchepin
    Danylo ShchepinEngineering
  • Khadim Fall
    Khadim FallEngineering
  • Valeriia Matkovska
    Valeriia MatkovskaDesign
  • Antonia Hartung
    Antonia HartungSupport & Operations

Two open roles

Engineer

Munich, hybrid.

The agent closes 35% of conversations today. Getting it to 80% is the most valuable thing anyone here could work on.

Mostly that means giving it more to work with. Today it reads the message text and little else. It should be reading attachments and log files too, and looking up a customer's real account state rather than inferring it. When the answer depends on something it can't see, it should ask a question instead of taking a guess.

What we look for

  • Several years building customer-facing products.
  • Strong React and TypeScript across the stack.
  • You measure before you optimize. If you can't tell us what got better and how you know, we're probably not a fit.
  • Enough product judgment to say when something we built is wrong.
  • Some curiosity about running LLMs in production. You don't need to have done it before.

Stack

React, TypeScript, Tailwind, Next.js, Node, Postgres, Redis. Zero for local-first sync. Tiptap for editing. Inngest for background jobs, on AWS. Linear, Slack and Raycast for everything else. You won't have used all of it, and that's fine.

How we moved our sync engine to Zero

Email Raphael about this role

Growth

Munich, hybrid. Base plus commission.

There's already traffic coming in. A few thousand people a quarter find us through Google, and the Linear integrations directory sends a constant stream of teams who work the way we do. Neither channel has had anyone on it full time.

Your first job would be working out which of those actually compounds, then putting real weight behind it.

You'd also sit in the sales calls. We charge per resolved ticket, so the customers worth winning are the ones with serious support volume, and companies like that don't arrive through a signup form. Someone has to talk to them.

What we look for

  • You can write. Send us something you wrote that you still stand behind.
  • You know the difference between a post people actually read and one that is correct and ignored.
  • You have opinions about software, and you can explain them to someone who disagrees.
  • You'd rather run one experiment properly than five badly.

One ask

Please don't send an AI-written application. We'll notice, and writing is the one skill this job really needs.

Email Raphael about this role

How hiring works here

The process is the same for both roles.

  1. 1A conversation with Raphael, the founder.
  2. 2A paid test task on something real from our backlog. We pay for your time either way.
  3. 3A longer conversation with the rest of the team.

Neither role fits?

Write to Raphael anyway, especially if you disagree with something above. [email protected]