Founding Engineer (m/w/d)

The opportunity

Tax advisory runs on structured data, recurring work and clear rules, and almost none of it has been rebuilt with modern software. What decides who gets there first is real case data and engineers who build next to the people doing the work.

Vogel advises several thousand mid-sized companies and acquires established advisory firms, developing them further on one shared base. On top of the systems we buy, we build our own software: a data layer above DATEV, the tools our advisors use every day, and agents that take over recurring work. What counts is daily use, on data that is audited and closed.

As one of our first engineers, you build across the whole stack: data pipelines, backend, the interfaces advisors use and the agents on top. You take a problem from the first conversation with an advisor to software that runs every day. We are hiring our CTO in parallel, who will own the architecture and lead the team. You work directly with Moritz Vogel.

What you'll do

Tools advisors use

• Sit with tax advisors and accountants on real cases, find where the work gets stuck, and turn it into software they use the next day. Your benchmark: an advisor would miss it if it were switched off.

• Build interfaces, dashboards and internal tools, from the first prototype to something that runs without you.

• Ship small and often: a tool in daily use beats a perfect plan.

Data and integration

• Build the pipelines that move data out of DATEV and our other systems into our own data layer every night, checked from the first day.

• Define every figure exactly once, so a number means the same thing in every tool and report.

• Connect the systems each acquired firm brings, so one more firm never means one more integration project.

Agents and automation

• Build agents that take over recurring work, with a clear line between what the model judges and what is calculated deterministically.

• Make every result traceable by version, source and decision, so a certified professional can sign off on it.

• Evaluate against closed, audited cases, not against test data.

Questions you'll answer

• Which recurring task in an accounting team is the first one worth taking off people's hands, and how do you prove it?

• How do you build fast now without leaving a mess for the next engineer?

• When is a prototype ready to run on real data every day, and who signs off before it does?

What success looks like

• 3 months: Your first tool is in daily use by an advisory team, and the nightly data load runs with checks.

• 6 months: Several tools and the first agents run in regular operation. You own one area end to end, and others build on your code.

• 12 months: You own a part of our software with its own users, and new engineers learn from your standards.

Where it leads: from engineer to owning a product area, as a senior engineer or by leading a team, decided by contribution, not by years served.

What you bring

• A degree in computer science, mathematics, physics, engineering or a related field from a strong university such as TUM, or an equally strong record of your own projects.

• You have shipped software that other people used, at work, as a working student or in your own project.

• You write production code in at least one language, ideally Python or TypeScript, and you are at home with SQL and relational databases.

• AI is your everyday tool: you write, review and test code with it, and you know where it goes wrong.

• You want breadth across data, backend, interface and agents, and you work in German and in person, because our advisors and clients do.

• You learn from practitioners, and you set the highest standard in everything you ship, not because someone checks, but because mediocrity bothers you.

A plus, not a requirement

• One to two years as an engineer in a young startup or scale-up, or a working-student role where you owned something real.

• Knowledge of accounting, tax or DATEV, or data pipelines and agents you have run in production.

• No conventional CV required: a mathematician or physicist who builds, an information systems graduate with a real engineering track, or the founder of a small product. We assess contribution, not labels. The range shown is not room for negotiation but a profile scale: the lower end stands for starting right after university, the upper third for someone who already carries the standard work of the role on their own. A different background starts at a different market price and converges on the role's level after twelve months of proven performance.

What this role is and what it isn't

• Building on a running business, not a greenfield. DATEV runs and clients do not wait. What you build has to work next to daily operations.

• An operating company, not a software vendor. The users are our own teams. What we build and in which order is decided together with management; you shape how.

• Breadth, not a specialist track. If you want to perfect one layer for ye