JOIN is looking for an Engineering Manager (m/f/d).
JOIN is an HR-tech company from Berlin that helps thousands of other companies find employees. Their mission is to make hiring faster, fairer, and simpler. It’s a platform that processes millions of candidate interactions. And now, ironically, the creators of a hiring tool themselves are looking for someone for one of the most “human” positions in IT — an Engineering Manager.
Let’s figure out what their pain points are. According to the description, the team has grown and become “mature.” This is a polite corporate way of saying: “We have a lot of smart engineers, but chaos has become more prevalent than order.” Processes are starting to creak, feature delivery is slowing down, and engineers need attention, a development vector, and someone to conduct One-on-Ones with them. The company needs a human glue that will bind the team, processes, and business goals, while maintaining quality and hiring new recruits. Classic growing pains.
But what if, instead of hiring an expensive manager in Berlin, with all the relocation packages and an M3 Pro MacBook, they invested part of that money into creating an “AI Co-pilot” for the team? Let’s fantasize about what that would look like.
Imagine a dialogue in a JOIN meeting room.
“We need a manager. A person who will feel the team, motivate, resolve conflicts. AI isn’t capable of that,” says the Product Director.
“It doesn’t need to be,” I reply, sipping my coffee. “It needs to become an invisible conductor that gives the team everything it needs to manage itself, allowing me or the team lead to spend not 20 hours a week on ‘human’ issues, but 5, with maximum focus.”
Here’s how it works in practice.
1. Task and Project Management. Forget about Jira, which everyone hates updating. We implement a system where an AI agent is connected to Slack, Git, and Jira. An engineer writes in the general channel: “Finished work on ticket AUTH-123, sending for review.” The AI agent automatically updates the status in Jira, pings the necessary reviewers, and, after analyzing code complexity, provides an approximate time estimate for the review. It also sees that an epic deadline is at risk, and doesn’t just color an icon red, but writes in the channel: “Team, according to my projections, we are 2 days late with feature X. The main bottleneck is testing component Y. I suggest Sarah and Max connect tomorrow and conduct pair testing.” This is no longer just a task tracker; it’s a proactive assistant.
2. Mentorship and Development. The most sensitive topic. AI won’t replace a heart-to-heart conversation. But it can prepare the ground for it. Before a One-on-One with a developer, I receive a summary from the AI: “In the last 2 weeks, Ivan closed 8 tasks, but his pull requests have been receiving more comments on code style. He also spent 30% less time coding than usual but actively participated in discussions about the new architecture. Possible topics for conversation: burnout, interest in an architect role, need for additional training on our guidelines.” I go into the meeting not empty-handed, but with data, and our conversation becomes significantly deeper. The AI can also analyze the code of all developers and suggest topics for internal meetups or recommend specific courses.
3. Hiring and Onboarding. Who better than JOIN to know how much time hiring takes? An AI agent can take on 80% of the routine: initial resume screening based on specified parameters, conducting automated code challenges, coordinating interviews. And for new employees, a personal AI mentor is created — a chatbot trained on all internal documentation. Instead of bothering colleagues with questions like “where is our…,” the newcomer asks the bot and gets an instant answer with a link to Confluence.
4. Reducing Distrust. Engineers are skeptical people. They won’t trust a “black box.” Therefore, implementation must be gradual and transparent. We start small: an AI that simply summarizes meetings and writes a daily digest of key events in Slack. Useful? Yes. Scary? No. Then we give it the ability to suggest, not command — “I recommend assigning Peter as a reviewer.” The team sees that the recommendations are adequate, and gradually, the level of trust grows. The key is that AI doesn’t command, but advises and automates routine tasks.
How to know if it worked? It’s very simple and measurable in numbers.
First – development metrics: Cycle Time (time from first commit to release to production), number of bugs reaching the user, time spent on code review. If the AI conductor is working, all these indicators should improve.
Second – anonymous team surveys. Questions like “How clearly do you understand the goals of the current sprint?”, “Do you feel you have time for professional development?”. Engineer happiness is quite measurable. If it grows, and with it the speed of delivering value to the business, then the system is working.
So, perhaps JOIN is not looking for a person for a director’s salary in Berlin, but an opportunity to invest in a system that will make their “mature” team truly autonomous and strong. This is a bold step that challenges the very essence of management. But it’s written in their values: Be Bold – Challenge the status quo.
Источник: https://www.linkedin.com/jobs/view/4408678656/
