[ninetyninenine]

In short, you manage people. You tell others what to do and you think about what to do.

It's technical management. You create impact off the work of others. You tell 10x people to do 10x work and you get 10x the credit when management only is like 1x of talent and effort.

I met a principle engineer who didn't know what a database transaction was and it still fits every description on this site. We shouldn't call these people engineers anymore. They are managers that don't need to do 1 on 1s.

[lumost]

There needs to be roles for pure ICs. In deep tech orgs, it’s almost impossible to get work done unless you have a bench of folks with domain expertise, comfortable on your software, and in tune with and guiding product direction. Diluting the core IC responsibility in these orgs with mandatory influence/scope guidelines is counter productive.

That being said, when I was in the role of a principal engineer at a major tech company - my ultimate responsibility was to get the project to success. What that meant on any given week could have alignment, coding, guiding, or reviewing.

[procaryote]

Some companies suck at promoting the right people, but a good principal engineer is the 10x person, or someone who makes several teams more effective by helping them figure out the bigger technical issues.

If you think management is like 1x talent and effort, you should try management for a bit.

[ninetyninenine]

I have. It's more stressful. But it's still 1x talent and 1x effort. Anyone can do it, it's just harder in the same way being a garbage man is hard.

10x engineers don't need to be principals. The idea behind principals is that they don't do any engineering. They leverage the work of others to become 10x. It's very much considered a "leadership" or management role overall.

A 10x engineer is usually literally a person who can output 10x through his own work alone. As a sidenote, with AI everyone is now 10x. So it's really a 20x thing now.

[procaryote]

I've seen zero people become 10x because of AI. Conservatively it seems closer to ±0.5x and it takes a while to figure out if it's + or -

I didn't understand the point of managers until I got a good one. Perhaps you've just been unlucky

[ninetyninenine]

I literally see the opposite. What I'm seeing is if you're not using AI, you're going to be left behind. What used to take a week or more is now expected to be finished in a day thanks to AI. The expectations have shifted by a lot.

I do know there are entire companies who have not bought into AI yet because the paradigm shift happened so quickly. The AI of 1 year ago could not make someone 10x. The AI of today can and if everyone knew about it, they would be using it.

>I didn't understand the point of managers until I got a good one. Perhaps you've just been unlucky

There is a point. The point is to set direction and to shield the engineer from politics. That is the point. Let me make it utterly clear: The role of leadership and management is REQUIRED. I never said otherwise.

My point, however, is to say that the primary skill involved in setting direction and shielding the engineer from politics and all this "leadership" stuff is this:

it is hard work, but it is not skilled work. The garbage man works hard, but anyone can do his job. There are certain people who do a shit job at being a garbage man and there are people who do a good job, but the overall job is considered unskilled labor. That is essentially what "leadership" is. And like the garbage man, the job is required.

The main difference between the garbage man and the leader is essentially perception. Because leadership is essentially a form of control, perception is controlled. That means leaders are perceived to be better and they are thus paid more but the on the ground truth is that what they do is not skilled work.

[untrust]

> The AI of today can and if everyone knew about it, they would be using it.

The AI of today absolutely does not add 0.5x, I'm using cursor and copilot and they still are usually just a fiddly tool which gets it right half the time. Anything complicated enough to need me to review its work takes longer through series of prompting and correcting its work than if I did it myself, and anything trivial enough for it to one shot its not saving me much time on anyways. All for a costly monthly subscription.

[ninetyninenine]

Unfortunately my on the ground experience is drastically different from you. I use Claude code though so that might be the difference.

But AI for me is right 90 percent of the time. It’s possible our prompt engineering is different.

> The AI of today absolutely does not add 0.5x, I'm using cursor and copilot and they still are us

You know when people say shit like this I wonder if they ever are able to think from another perspective? Like they say one thing but tons and tons of other people are saying another thing and what gets me absolutely curious is how someone can be so brain fucking dead that they can’t even consider the other perspective. Tons and tons of people say what you say but an equal amount say the opposite.

[procaryote]

This is a bit of a naive view of what a manager does. I've hired lots of managers and have interviewed a lot more, and there are definite real skills to the role. Those skills include hiring good people, fixing teams that don't work well, etc

It would be lovely if all they needed to do was point vaguely in some direction and stop other managers from talking to the team. I could hire much cheaper people for that

[ninetyninenine]

Those are skills those are some guidelines that once people follow it most people can do it.

Nowhere did I say we needed to point in a general direction and let it go.

[willseth]

Maybe you were just bad at management and didn’t know it.

[light_hue_1]

> In short, you manage people. You tell others what to do and you think about what to do.

As a principal scientist I definitely don't manage people.

Mainly, people managers have power to tell people what to do and when. I have zero power over anyone. No one needs to listen to me. Ever. If people ignore me there's literally no one I can complain to. And no one tells me what to do. If I wasn't competent at my job I'd be sitting alone in an office staring at a wall.

The reason why anyone would ever talk to me is because I have well over a decade of experience more than every other scientist in my org (200+ people) doing very complicated things, at scale, across many areas of AI/ML, and publishing many dozens of papers in top venues. So I know what I'm doing and can find problems, shortcuts, keep them from wasting time, find creative fixes to AI/ML issues, and see connections to business problems. Basically help scientists deliver better features faster. I can find new projects that those scientists are excited about. And I can communicate with product, managers, leaders about very complex technical ideas in ways that more junior scientists cannot.

People in my org come to me because I provide value basically. Not because they have to.

> You tell 10x people to do 10x work and you get 10x the credit when management only is like 1x of talent and effort.

If only! 70% of the work I do is invisible small course corrections here and there across multiple orgs that fix things no one ever hears about but would be disasters down the road. 20% is working on projects that scale with many people. 10% is strategy for new experiments and projects.

For the vast majority of what I do other people get 100% of the credit and my name is a small footnote at best.

Maybe an example would help. Sorry that I need to be a bit vague. A large project recently was going in a particular direction. I saw two months ahead of time that this direction was going to run out of steam for a combination of science and business reasons. It would end up missing metrics. Leadership was asking for something that was broadly correct but subtly wrong. I spent a month finding a new direction that was the opposite of what everyone thought we had to do, finding the right small 1-2 person experiments to derisk the new direction, get senior people to understand it and agree to it, set up the correct relationships to make it viable, and then I had the evidence to convince project leadership and management to shift. A few days later we had a VP review and the response was that they're delighted with the direction, it's far better than they ever though, and this will be a flagship for an org with tens of thousands of employees. That review would have gone horribly without what I did because the problem I saw two months ahead of time would have surfaced and the project would be seen as dying and aimless. But the team got the credit for being on top of things as a whole, not me. As it should be.

> They are managers that don't need to do 1 on 1s.

I have more 1:1s than managers, because they can afford to just tell people what to do. I need to develop junior scientists, keep relationships up, stay on top of other orgs, business priorities, etc.

Managers fail by creating chaos around them. Principals fail by becoming irrelevant.

[neilv]

This is much like real Principal Engineer roles I've seen.

The skeptics of the role might not have seen how easily and badly a large multiple-team effort can go astray, and the value of someone spotting the problems, and making sure they get addressed, before the product line or company is ruined.

[ninetyninenine]

I agree with your second sentence. But Anyone can do that, it's not a skill exclusive to principals. The idea that only certain types of people have the intelligence to spot things like this are principals is wrong.

[neilv]

Agreed, many people do that, and that's great.

Maybe we should think of the Principal role as complementary? If you're working on a compiler, you're going to be interacting cross-team with hardware engineering team, etc., and spotting lots of things. But someone who is looking at all the teams, and not spending so much time on compiler details specifically, will spot some things the compiler person doesn't. So together you get better coverage of problems that neither alone could spot.

Of course, once we throw differing pay grades into an organization, everything gets more complicated. And people might resent something being called "complementary", if what's bothering them is that the role in question pays better or is considered more prestigious.

Though, the Principal role is in the engineering career track of that compiler writer, if they want that kind of ulcers.

[ninetyninenine]

it can be, but the fact of the matter is the role is viewed as "better". Principals are better, smarter, more experienced and higher ranking than people who don't hold that position.

[sandeepkd]

From what little I have seen, this kind of role is tightly coupled and dependent on the their Manager. The manager has to you like as a person and some how believe that all these activities are adding value.

[light_hue_1]

> From what little I have seen, this kind of role is tightly coupled and dependent on the their Manager.

As you go up the chart you have more independence and are less tightly coupled to your manager. By the time you get to principal you should be largely independent. At the same time, you have much more responsibility.

That's just a practical problem. As your manager becomes more senior (director/VP) their scope also increases. They just cannot "manage" you the way someone would manage a more junior IC. Also at the principal level you aren't just bringing value to your manager, but to other parts of the org as well.

In other words, I can't ask my manager "what should I do today?". I cannot even imagine what his reaction would be if I asked that question.

> The manager has to you like as a person and some how believe that all these activities are adding value.

For what it's worth my manager is a great person. But he wouldn't for a moment believe anyone when they say they add value.

It's up to me to find ways to document and express my value. Figuring out how to do this is part of becoming a principal. So I keep notes, I record wins, I make sure that I do things that bring me visibility, that I present new ideas, I contribute to larger roadmaps at the org level, I make sure that other scientists can say good things about me, I help fix problems that other orgs have so that they report I was useful, etc.

[re-thc]

> People in my org come to me because I provide value basically. Not because they have to.

In most orgs sadly that's not why people talk to a principal. It's that the process requires it, the manager wanted them to and/or there's you i.e. someone else to take responsibility. Of course they won't tell you that directly.

[taneq]

How do you get a role like this (even with the 10 year head start) without owning the organisation? Very few places I’ve worked have the kind of long-horizon pragmatism to invest in this kind of role.

[trashtensor]

Big companies tend to have Staff/Principal/Distinguished type roles. Usually those roles give you a lot of independence. But that independence means you need to be able to find projects to do and advocate for them and get them staffed and planned and executed and out to production. Often those projects can span many teams and multiple organizations, depending on how the company is structured. So I suppose the most valuable skill is being able to earn the trust of the managers so that you're able to even get them to listen to you so your stuff ends up on their roadmap.

so i do sympathize with a lot of the negative sentiments about the role here in this thread, and i think that in general there is a lot of navel gazing about the staff+ tech ic roles, there is an actual place for them as tech leads of large projects.

[questionableans]

One way to gather the skill set is to rotate through and master different aspects of the role. First become really solid at the technical fundamentals. Then volunteer to spend some time focusing on complementary skills like project and product management (in very technical areas, where you couldn’t just hire a nontechnical person in one of those job families). Start looking out for the problems worth solving that take multiple technical experts to accomplish, and make the solution great. Maybe be a manager for a while. Understand how to make your managers for at least a couple of levels up better in some ways, and help them with that. You will become someone trusted to get the right things done well. When it comes time to find the next job, focus on carefully vetting opportunities and interviewing prospective employers at least as much as they do when hiring, to find that opportunity you can be truly great at.

[procaryote]

If you have the skill and you're in a smaller company, you can just keep doing the role and the org will adapt to use you in it

If you have the skill and you're in a larger company they may already have a structure around it you can apply. Otherwise you should talk your manager into trying it, or you should find a new job in an org that does have this kind of role

If you don't have the skill, you should get the skill first, by making your team more effective, then make your team and all related teams more effective together

[ninetyninenine]

Politics. It's less about skill and more about social engineering.

[ninetyninenine]

>But the team got the credit for being on top of things as a whole, not me. As it should be. >If only! 70% of the work I do is invisible small course corrections here and there across multiple orgs that fix things no one ever hears about but would be disasters down the road. >For the vast majority of what I do other people get 100% of the credit and my name is a small footnote at best.

Then your case is the exception, not the rule. To reach the level of principal, you generally have to be recognized for delivering high impact. Visibility is not just ego, it is how organizations perceive value. If your work earns you no visible credit, then no one really knows what you contribute. And if no one knows, how could anyone justify promoting you in the first place.

That is the ideal. In the ideal world, principals are elevated because they have a visible history of making the system better. They build frameworks that others rely on. They turn chaos into structure. They guide teams through impossible projects. Their reputation is not something they chase, it forms naturally from the wake of their work. In the ideal, visibility is the residue of real impact. People talk about them because their fingerprints are on every success.

But the corporate world rarely functions on ideals. In the real world, power accrues to whoever is closest to power. Titles often flow through social gravity more than technical merit. Some people climb because they deliver, others because they simply survive long enough to become unmovable. The higher you go, the more politics matters and the less evidence is required. Impact becomes subjective. Influence becomes reputation. And reputation, once earned, decays slowly.

In that reality, being invisible is not a liability. It can be a strategy. A principal who keeps their head down, avoids controversy, and stays on friendly terms with the right directors can outlast a dozen brilliant but abrasive engineers. The irrelevant survive because they are not a threat to anyone’s ego. The company quietly carries them, paying tribute to their title while forgetting their function.

Even the ideal, though, cannot escape the need for visibility. A principal who does great work in secret still fails the fundamental requirement of leadership: to be seen. Influence requires perception. You cannot guide a culture if nobody knows you are there. Quiet impact might keep systems healthy, but it does not create belief, and belief is what organizations promote. The best engineers learn to make their results legible. They translate their impact into stories others can tell. Without that, the work disappears into the background noise of everyone else’s effort.

So there are really two systems running in parallel. The first is the ideal, where promotion is earned through visible excellence and quiet authority. It demands both impact and awareness. The second is the reality, where promotion is often granted through time served, connections maintained, and an ability to avoid friction. The ideal rewards contribution; the reality rewards endurance.

You can succeed in either system, but they ask for different currencies. The ideal asks for mastery, courage, and the discipline to lead by example. The reality asks for patience, diplomacy, and the instinct to stay useful enough but never threatening. One builds respect. The other builds stability.

And most companies, if we are honest, prefer stability. Stability requires engineers to act the way you describe but talk the way I do. You talk about the ideal, you even believe you walk it, but because no one can see your impact, no one can tell the difference.

I've held the ranking of staff in many companies. I've interacted with principals, staff, and distinguished engineers and I can tell you visibility is required to fulfill the ideal. If visibility wasn't there, than the person earned the rank through other un-ideal means.

[light_hue_1]

> Then your case is the exception, not the rule. To reach the level of principal, you generally have to be recognized for delivering high impact. Visibility is not just ego, it is how organizations perceive value.

Of course you need visibility and impact. But it's not the kind of crass taking credit for what other people did that the original poster talked about.

As a principal you need to actively build visibility because so much of your work is normally invisible. It means being an active participant with discussions with leadership, clearly being the person that sets the agenda on something, becoming someone who other principals turn to on a topic and then report to their managers, leading the conversations on a topic, being the person that people can bring a hard problem to and then see results, picking up on a business need first and finding a way to deliver on it, etc.

In the example I gave for the quiet change in direction of that project, there are many ways in which I get visibility. I told my manager this was a risk and I have a strange solution. I told the manager of that project. The other principals that I needed to involve and their directors know I pushed this. My VP knows because no one talked about that kind of feature until I did. I checked how crazy this direction change would be with more senior scientists and product people. I actively make sure that these gentle communications happen, and in a sense they're natural, because I'm taking the lead on something.

But no one is going out of their way to say "I take credit for all of this work that everyone else did".

> A principal who keeps their head down, avoids controversy, and stays on friendly terms with the right directors can outlast a dozen brilliant but abrasive engineers.

I don't think that abrasive people should become principals until they change their ways. There's so much cross-org coordination that you need to do, if you're abrasive, that's going to hurt everyone.

That doesn't mean you should be a pushover. I don't back down from technical arguments if I know I'm right and have the data to back it up. I have gotten into deep weeks-long disagreements reported far up to chain. But it's important that you can still go have lunch with your peers and collaborate on other topics even while you're trying to show that they're totally wrong in one area. That's part of building trust.

> In that reality, being invisible is not a liability. It can be a strategy.

A strategy to be fired. This is not viable.

There is no tension between "quiet authority" and visibility. If you're invisible no one will ever come to you. Visibility is what consistently demonstrates to people that talking to you will make their lives better.

[ninetyninenine]

>A strategy to be fired. This is not viable.

Not true at all. Plenty of invisible people in the world who don't get fired. You're ignorant.

Your initial post talked about lack of visibility, now you're talking about it as if it's required comically to the point of claiming that "invisible" people are absolutely on the path to get fired or that people who keep their head low and don't grab attention isn't a strategy. Ever heard of the saying the tallest blade of grass is the first to be cut? Also have you seen how Xi, the current leader of China rose to power?

Part of being a principal is the perception of principled and expert reasoning/expertise. That means being able to flip what you're saying on the fly so that other people continue to perceive you as an expert who knows what they are talking about. Did you say something that was completely off and wrong and it made no sense? How do you hide this from the people above you?

You expertly did this with your response. First you say you take none of the credit, now you say taking credit and being visible is required. Expert pivot and THIS is truly the primary skill of a principal engineer and you display it unequivocally.

You don't even need to actually be responsible for all the changes you claimed to have influenced. You just need people to perceive the reality as if you were responsible. And I see this kind of BS in many, many engineering organizations.

A lot of this is self delusion too. These high level strategies and directions aren't hard to come up with. Likely tons of lower engineers have thought of the method too, they just don't have the "visibility" or the time to actually drive that direction into the organization. Some people tend to think the ideas they have are genius and that these ideas are what makes them "principle". No. It's mostly politics that puts you into a position to take credit for ideas anyone can come up with.

[DiscourseFan]

I guarantee you that the "10x" engineer would not be able to co-ordinate anything with any other team or department in the company. Of course the hierarchical structure of companies means that just having good people skills, and good management skills, means better compensation. But the most important skill in any capitalist economy is organizing labor, since capitalism is a way of organizing labor, so those who are the best at organizing and directing labor are the most well compensated since they are able to produce the most profits.

[BobbyJo]

The most competent engineer I've ever worked with, the only one I'd consider 5x+, was fantastic at cross team collaboration.

[microflash]

This is just another stereotype. I’ve seen plenty of managers struggling to form coherent communication where an engineer would express with perfect ease. The problem with communication comes when you add layers of bureaucracy between people, often under the guise that engineers can’t communicate.