All positions
Engineering Vollzeit Hybrid, Hamburg office €5,000–6,000 fixed / month

AI Software Engineer (FTE)

What it’s about

You build the platform on which agents do sales and marketing work. It is our own product: tables that research and score companies and contacts. A sequencer that guides leads across email, LinkedIn and WhatsApp. Workflows that react to events. Agents that make the judgement calls a fixed rule cannot. And the integrations that connect all of it to CRMs, mailboxes and data sources.

You work close to the founding team and see directly what your work does in production: the system also runs our own sales. When something does not work, we notice the same day.

CegTec builds software for GTM with forward deployed engineers at its side. Profitable, bootstrapped, growing.

How we work

Every role at CegTec works with AI agents on the same principle:

  1. Set the goal: You define what should be achieved, for example 20 qualified meetings a month.
  2. Agents build: Agents do the work: research, data, drafts, routine.
  3. Evaluate the result: You review what comes out and give clear feedback: good, not good, and why.
  4. Iterate: The agents get better with every round until the goal is met.

That way your time goes where it matters: talking to people, meaning prospects, customers, partners and the team.

What you own

  • Platform features and operations — from the data model to the interface, from the first idea to stable operation.
  • Agents — tasks, tools and checks, so agents work reliably and their output is measurable.
  • Tables and sequencer — the execution layer: columns, gates, send windows, unsubscribes, idempotency.
  • Integrations — CRMs, mailboxes, LinkedIn, enrichment providers and third-party APIs.
  • Quality — tests, observability and clean migrations. Stability over speed where you cannot have both.

What you explicitly don’t do

You do not run sales and you do not run customer projects. Customer work sits with the forward deployed engineers, sales with the founding team. You build the product both work on.

Who you probably are

One of three stories fits:

  • You are a software engineer focused on web and backend, and you want to work on a product where AI agents are the core, not an add-on.
  • You have been through startups or product teams, know the path from idea to live system, and want to work close to product and founders again.
  • You moved up early, built your own agent projects on the side, and want to do it full time.

Requirements

  • Confident in TypeScript and the web stack: Next.js or React, Node, Postgres/Supabase
  • Experience with LLMs and agents: tool calling, prompting, evaluating results
  • Clean work on production systems: Git, tests, migrations, code review
  • Daily work with AI coding tools (Claude Code or comparable)
  • German solid, English good for code and documentation
  • Living in or near Hamburg, regular presence in the office

Bonus

  • Experience with APIs of CRM, email or LinkedIn providers
  • Experience with queues, background jobs and idempotent flows
  • Own projects or open source we can look at

What we offer

  • Permanent position at CegTec GmbH, hybrid: Hamburg office + remote, living in or near Hamburg required
  • €5,000–6,000 fixed per month, depending on experience
  • Direct collaboration with the founding team, no middle management
  • Premium AI coding stack (Claude Code, GitHub, all cloud accounts)
  • You work on a product we use ourselves every day

Selection process

Four steps over 3–4 weeks: introductory call · walkthrough of something you built · practical task on a real case · references and offer.

How to apply

Send your pitch by email to jobs@cegtec.net or via LinkedIn to Luca Ceglie. Bring two things:

  • Three sentences — which of the three stories fits, what draws you to the product, what you would touch first.
  • Something you built — a repo, an agent, a service. Show what it did and what you changed because of it. A link, a screenshot or a page of text.

No cover letters.

Application

Apply now

Two quick questions before we begin.

Question 1 of 5

Have you written production code in TypeScript (or a comparable language) that real users used?