When you advertise for an Engineering Manager, are you sure which job you mean?
When you advertise for an Engineering Manager, are you sure which job you mean?
Over the last few months I have been talking to a lot of companies about how their engineering teams are run, and reading a lot of the EM ads going out. It is becoming clear the same title now covers two different jobs. Some of that is AI, some of it is just the profession growing up. If you are about to hire an EM, or restructure around one, it is worth taking a moment to decide which one your team needs.
The UK market is selective, so your ad needs to be specific
Hiring has been slow for nearly three years, and the latest KPMG/REC reports say it is only now stabilising. Most roles going are replacements, not growth. From the hiring side the numbers are hard reading too:
- One hiring manager told The Pragmatic Engineer they get around 1,000 applications a day, and maybe two are relevant
- Ghosting is more common in the UK and EU than in the US, on both sides of the table
- Fully remote roles are shrinking, and some work is moving offshore
- The hiring pipeline for management roles is down 43% since 2022
The good news is that a clear ad cuts through that noise. The companies I see hiring well are the ones who know exactly which EM they need, and say so.
The role is splitting in two
The EM job is going two ways:
Tech lead manager. A small team of three to five. They still write and review code, set the technical direction and run the one-to-ones.
Multi-team EM. Several teams, wider scope, mostly people, planning and reporting up. Running one-to-ones on work they don't know deeply.
Think of it as a factory floor. One is the team leader who still runs a machine on the line and knows every setting. The other is the production manager walking between lines, watching flow and handling the people side. Both are needed, but they are not the same job, and you would not hire one to do the other.
Most companies have not chosen a model on purpose. It has happened through restructures and layoffs, so the job ad still reads like the older role. Then you hire a production manager when the line needed a team leader, or the other way round, and it takes six months to find out.
AI took the status reports, not the judgement
A big part of why this is happening now is AI. Status updates, progress tracking, meeting notes and scheduling are being automated. That was a lot of the coordination layer EMs used to carry, so it is tempting to think you need fewer of them.
In the most recent team I advised I saw this change happen. The tools picked up the admin, and the team moved faster on some things. But people still needed someone to keep them together and on track. Someone to notice when a pair are pulling in different directions, when a ticket has stalled because nobody wants to own it, or when a good engineer is not sharing updates in meetings.
AI also adds work. There is a lot more generated code to review, and someone has to decide if it is right for the system and not just right for the test. That is a technical judgement call, and it is where the hands-on manager earns their place.
What did change was how work came into the team. The PM and EM could lean in earlier, improve the PRD and the research, and get more decisions signed off with stakeholders up front. It is a slightly more Waterfall approach, and it suits senior management, as progress updates are more predictable for them. The skills and guardrails get refined into the process, so the team works more like a hive mind on the project.
Though the task structure is more predictable, the output and PR sizes are still a bit like throwing a refined dice. Sometimes it needs picking up and putting back on the table. Tip here is to give the PM access to snapshot builds for frontend work (Cloudflare Worker Previews can spin up an instance per branch). Then the PM can test and review progress, and the EM can challenge the engineer to explain and refine direction. That moved us away from "builds on my machine" and diff-only PRs that were hard to reason about in mixed status meetings.
These are close to soft skills though, and not everyone uses them the same way or can decide which way is right. So we still needed an EM with oversight of the projects in their domain, hands-on enough to lead the refinement, and to decide when to draw the line or push a deadline.
Also, not all engineers are fully acclimatised to being AI native yet. Even the experienced ones are only a year or so into this change. Building that trust and growth alongside their professional competence still takes a manager.
Answer these before you write the ad
Tip here is to answer the questions a good candidate will ask you in the interview, before the ad goes out:
- How many engineers will report to this EM, and has that changed in the last 12 months? (Past 10 to 1 you are in multi-team territory)
- Do you expect them to ship code or review pull requests? When did your last EM do either?
- Who reviews AI-generated pull requests now, and who decides what good looks like?
- Is the gap a "lead engineer" or "player-coach", or a "manager of managers"?
- What did your last EM promotion look like, a strong engineer on one team or someone growing across teams?
- What should they own in the first 90 days, one team's delivery or a set of teams' process?
None of these are trick questions. If you have the answers ready, the right people will recognise the job, and it saves both of you months of the wrong fit.
Hire for the pattern you need
If you are building something new, the tech lead manager earns their keep early.
Back in 2018 on a new SaaS build, myself and the senior developer spent four weeks building and testing the core services of the MVP before anyone else joined. Scaling out to two more developers was then easier, as the next ten features could follow and build on those architecture patterns. If four developers had started at once we would have had four slightly different patterns, and paid to unpick them later.
With agentic programming this pattern is still around, but the time frames have changed. The early work is now plans, skills and guardrails, plus a few first "golden references" that flesh out the architecture. Human-readable PRDs and models for future services keep the direction a collaboration between AI agents and the engineers leading different parts. Someone has to own that, and it is the hands-on EM.
If you have several teams with different stakeholders, you need the other kind.
With so many applications for every role, an ad that says "we need both" reads as "we're not sure". Pick one and say it plainly.
Which EM is your team actually missing?
Take a moment before the next ad goes out to run through the questions above with your CTO or head of engineering. It might change the role you advertise, and who applies.
>
Not sure which EM your team needs?
Let's talk it through, I'm happy to look at how your engineering team runs today and where the gap really is.