Coaching & Mentoring Engineers
The senior engineer's hardest promotion isn't to staff - it's from being the person with the answers to the person who grows the people who have them.
For years, the job rewarded one thing above all: having the answer fastest. The gnarly production bug, the design nobody else could see, the migration everyone was afraid of - you were the one who knew. Promotion, respect, and the quiet dopamine of solving hard problems all trained the same reflex: I know, let me show you.
Then you get a mentee, or a team, and that reflex quietly sabotages you. Because every time you hand someone the answer, you rob them of the rep they needed to grow their own. The thing that made you valuable as an individual contributor is the exact thing that makes you a bottleneck as a leader. Nobody warns you about this. The skills do not transfer - they invert.
Four hats, worn by reflex or on purpose
The root confusion is treating "helping people" as one activity. It is at least four, and they only look alike from the outside.
You teach when there is a genuine knowledge gap - a tool, a concept, a piece of syntax they have simply never met. You mentor when you share hard-won experience: how promotion actually works here, what you learned the painful way. You coach when the person is capable but needs to find their own path, so you get out of the way and draw their thinking out. And you manage when the conversation is about org priorities, performance, and accountability.
Teaching and mentoring are additive - you put something into the person. Coaching is subtractive - you remove yourself so their competence can surface. Managing is directive - it carries authority and consequences. Most real conversations move between all four. A single 1:1 might open in coaching mode ("What's on your mind?"), pivot to mentoring ("Here's how the staff promotion really works"), and close in managing mode ("So the priority this sprint is the migration"). The failure is not picking the wrong hat once. The failure is never noticing which hat you are wearing.
The advice monster lives in your chest
The hardest mode is coaching, because it asks you to do less, and doing less feels like failing. Watch the difference in a single exchange. An engineer says the deploy is failing and they have no idea why. The advice reflex fires instantly: "Oh, that's almost always the env var, check DATABASE_URL." The coaching reflex asks instead: "What have you ruled out so far?"
The first answer fixes today's deploy. The second builds an engineer who can fix the next ten without you. Michael Bungay Stanier calls the urge to jump in the advice monster, and it is loud precisely because your identity is built on knowing. Silence feels like incompetence, so you fill it. The cure is almost embarrassingly simple: after they finish, count two seconds, and ask one more question instead of offering one answer. "And what else?" is the most useful question in the world, because the first answer is almost never the whole answer.
This is not a personality you either have or lack. It is a handful of learnable behaviours - listening past the gap where you would normally start talking, asking open questions that hide no advice, and running a loose structure like GROW: get the goal, get honest about reality, generate real options, then commit to a first step. Practised on purpose, it makes the people around you measurably better at thinking. That is the entire job.
Mentoring is not the same as sponsoring
There is a quieter failure that even good coaches commit. They give generously - advice, time, careful feedback - and stop there. But mentoring happens in the room with the person, and the things that actually change a career happen in rooms they are not in. Mentorship helps someone get ready. Sponsorship is staking your own reputation to say "she should lead this project, I'll vouch for her" when she is not there to argue for herself.
The two get conflated, and the cost falls unevenly. Underrepresented engineers in particular tend to be over-mentored and under-sponsored - drowning in advice, starved of the stretch assignment and the visibility that advice cannot buy. As you grow in seniority, your capital becomes more valuable than your calendar. Spend it. The most generous thing a senior person can do is not another hour of coffee; it is naming someone in a room where it counts.
Feedback is a frequency problem
Most feedback fails not because it is wrong but because it is rare. Saved up for review season, it lands like an ambush. The fix is to make it small, specific, and constant, and to strip out the character judgments that make people defensive. Situation, behaviour, impact - describe the specific moment, what a camera would have seen, and the effect it had. "You're dismissive in meetings" is a verdict. "When Sam raised the caching idea you moved on before he finished, and he went quiet for the rest of the meeting" is something a person can actually do something about.
And caring about someone obligates you to tell them the truth. Kim Scott's name for the trap kind engineers fall into is ruinous empathy - being so nice you withhold what they need to grow. Silence is not kindness.
The real test is the empty room
Coaching one person is hard. Building a team that grows whether or not you are present is the actual work, and it cuts against every heroic instinct you have. If everyone's progress routes through you, you have built a single point of failure, not a culture. The honest measure of whether you are succeeding is brutally simple: are people doing things now that used to require you? If yes, you are coaching. If they are merely dependent on you in shiny new ways, you are rescuing - and rescuing always feels like helping right up until you go on vacation and everything stops.
The promotion nobody hands you is the one from answer-haver to people-grower. You take it yourself, one swallowed answer and one better question at a time.
No comments:
Post a Comment