How do you manage up when your startup CEO is smart but inexperienced?
Posted by kuncog@reddit | ExperiencedDevs | View on Reddit | 46 comments
I’m at a small startup (~15 people) and acting as a sort of head of engineering: doing eng management, product shaping, code reviews etc.
Our CEO is head of sales and kind of moonlights as CTO. He has a CS degree and can code, but he’s a career founder who’s never worked at an actual company. So while he can build MVPs (he cranked out two with AI tools and got some early customers), the codebases are a mess: 5k line files of JS, no tests, no structure. Basically tech debt from day one.
We could clean it up, but priorities shift every week, so the eng team never gets a chance to do real work or planning.
Our CEO is on the younger side: he’s about 15 years younger than me (I’m late 30s) and not hard to work with, he listens, but he’s easily swayed by whoever he last talked to.
I’ve managed teams before, but never in a startup this early or chaotic. How do I help bring some direction that actually sticks?
bentreflection@reddit
You need to explain why it’s better to do things your way vs cutting corners to get things to market faster. But also, as someone who has spent my whole career at startups, sometimes (maybe even often) getting something out to meet current market demands is better than doing things the “right” way.
You need to be clear about when and why it’s NOT beneficial to accumulate more tech debt
kuncog@reddit (OP)
I’ve worked on MVPs in the past as well, I’m by no means advocating for enterpise grade software at this stage, however the AI slop that we’ve been working with slows us down unlike any legacy code I’ve seen in my career.
bentreflection@reddit
i kind of missed the part where your CEO is coding via AI and making you guys work on it. That's not great. If you can I'd try to have a discussion with him about having him not code at all. If he really wants to act as CTO you can have him guide general technical vision but he should definitely not be cranking out AI slop and making you guys deal with it. I'd just tell him straight up that it's slowing down velocity more than it helps.
I've worked at a couple startups with similar issues of the CEO trying to keep their street cred by coding and things always improved once they were told to stop doing that and focus on being a CEO. Too many times coder CEOs are just trying to code because it feels productive and being a CEO is nebulous and not as fun.
so yeah step 1 is to keep him away from directly working on the codebase. Step 2 i'd say is beginning to add tests for any new functionality going forward as part of the pull request / merge cycle. Step 3 will depend a lot on budget and team size but you could allocate some amount of developer time each week to addressing tech debt.
At some point technical debt will prevent you from adding any new features. As you mentioned already, if you're starting from the ground up with AI slop that technical debt barrier is going to happen a lot sooner.
kuncog@reddit (OP)
Yeah the big issue is the AI generated garbage code that technically works but actually who knows.
Good points about getting him away from coding, I will look to push him in that direction.
sciencewarrior@reddit
You can turn AI to your advantage. A couple of hours with Claude Code or Gemini CLI can give you a pretty decent test suite, in my experience. You can also run linters on commit or give Copilot guidelines to review new code.
As for how to tell to him, you can tactfully remind him that every hour he spends vibe coding is one hour that he doesn't talk to customers and potential customers to understand how they are using the product and what gaps they see.
Historical_Emu_3032@reddit
In a similar situation. New CTO brings a lot to the table but not in the way we expected.
But we need to learn how to guide him because his strengths are things we still need.
It's frustrating cause the c levels are your bosses but specially in a startup situation it doesn't mean they're the most knowledgeable.
unlucky_bit_flip@reddit
Your first mistake is trying to operate a startup like a mid-size or large company. And I’ve seen too much process early on, kill startups. (same with rewriting software)
Speed is everything. You won’t have much time for guardrails and “continuous improvement” because you have 6-12 month window before you run out of cash.
The best you can do is help prioritize things that move the business closer to financial success. The customer doesn’t give a fk about your tech debt.
kuncog@reddit (OP)
We have runway for at least 2 years, we are not that pressed for time.
I also get that speed is everything, but tech debt can and does slow things down.
vineetr@reddit
Others have pointed this out to you, but I don't think you see yourself as part of the problem.
20-year-old guy with a CEO title isn't going to figure out how to run a tech org. If you are the senior most engineer in a 0-1 startup, your role isn't to explain why your CEO's code is bad, how many test cases are needed to get high confidence about shipping to production, or anything else that is remotely technical.
It sounds a lot like you want decisions made for you. You want the CEO to write tests. You want the code to have structure. And you want someone else to define the engineering processes and structure for you. That is the problem. That's the lack of agency the other comment was alluding to. Sure, the CEO is creating the environment for you to fail with his antics, but the lack of assertiveness and ownership is rather troublesome.
If you were expecting someone to agree with you on the need to create processes etc. you've missed the elephant in the room - this company is a startup. It probably doesn't have PMF - a few paying customers doesn't imply PMF. 2 years of runway does not imply PMF either. The 2-year figure can very easily become a smaller number with some bad decisions.
What you should be doing, is figuring out what need to be shipped in those 2 years, how soon you should be shipping, and define just the minimum amount of structure for that. If you are in the phase where you are making POCs to close deals, then it is perfectly acceptable for those POCs to have tech debt on day 1. The money hasn't come in for tech debt to be addressed. You can fix that debt, once there is money to fix it. If a feature isn't making money, spending more money to fix something you don't like about how it is written, is poor judgment. That's how technical people lose credibility in front of business folks.
bill_1992@reddit
Having worked with founders, almost every one of them knows what tech debt is and wants to eliminate it.
In a traditional mid/large tech company, you make a big deal about something like tech debt, you get the VPs approval to allocate resources for it, then you get promoted when it's all done. And it's rewarded because that's how larger companies operate. However, that's not how startups work.
What startups want you as the technical leader to do is to take care of the problem. They don't want to hear about it because it's not really their area of expertise and it doesn't directly impact success in any way, but they'll often give you the leeway to fix it on your own - that is the value prop of working at a startup. If you truly are a good engineering leader, then you should be able to manage tech debt and set an example for other engineers without compromising the business.
It's really clear in your comments that you have absolutely no respect for your CEO, but quite honestly, I think you made no attempts at introspection because you've just default blamed everything on your CEO. You're not looking for advice here, you're just looking for validation.
Designer-Relative-67@reddit
The ceo is in their early 20s, i think its fair to say they probably are the problem
Howl33333@reddit
Right. Most sensible comment here - you’re working at a startup. Own up some agency and just build. Do both. That’s what you signed up for.
Desperate_Rub4499@reddit
how honest can you be with this guy without him hating you?
db_peligro@reddit
hopefully this guy is self aware enough to know his limitations and accept help from more experienced people. if the guy is arrogant there's no nothing you can do.
kuncog@reddit (OP)
He is not arrogant, the opposite in fact, but I’ve found that he is easily influenced.
For instance, I gave him the spiel about having some sort of direction, backlog, at least a 2-3 month plan but then a random investor mentioned to him that “startups need to pivot fast” and that essentially meant we throw any medium term planning out the window.
SarriPleaseHurry@reddit
I mean you aren't wrong. And neither is the investor. He has no company if he doesn't listen to the people funding him. You have no job if the company doesn't pivot on something that is a sink cost.
Does the company have PMF? Is it profitable? Depending on what stage the company is in pivoting quickly matters way more than code quality and long term planning.
kuncog@reddit (OP)
Yes on PMF, no we’re not yet profitable but we are making money and getting some real interest from clients. Runway should last for 2 years easily and we should be profitable by then (per projections).
Also, I’m not looking for long term plans (>6 months). Having a 1 month plan at this point would be already a win.
gomihako_@reddit
It sounds more like you're the CTO
TwoPhotons@reddit
As someone who was in a similar situation the hardest part was getting them to listen to you. Often they would just nod along and agree with everything you say, only to continue doing what they were doing before. However in my case I was younger than the CTO so I was at a disadvantage from the outset.
kuncog@reddit (OP)
This actually matches my experience, always in agreement then he changes his mind a couple of days later.
kuncog@reddit (OP)
Not too honest, it’s his company, he is the money guy and he gives me the impression that he rarely heard the word “no” while growing up.
Desperate_Rub4499@reddit
id say adapt to ur environment. what is an actual problem that exists at the current maturity of the business. dont sell him solutions as if its a enterprise software if it isnt. since its small startup, just take initiative and start doing stuff and making it better. find his weak points and then adapt.
if you’re “head of engineering” u should be building systems that work for the stage of the business and enforcing them. he hired you to be a lead right?
but in some cases at startups, the mvp is fine and even though the code is shit that doesn’t matter unless ur trying to pitch like scalable architecture to VC. dont over systemize and pretend u have spotify level traffic right now basically if u should just be helping people simply not break prod every other day or asking for env files. whatever solves todays little problem
calloutyourstupidity@reddit
gg wp
kuncog@reddit (OP)
Yeah I mean some people are born with a silver spoon in their mouths and then the rest of us are lucky to work for them …
TedW@reddit
Was that simpler than saying early 20's?
g_bleezy@reddit
“How are we going to measure success?” Ask for clarification, specific numbers, timelines for those results, get really fucking specific. Over and over and over again. Get it in writing, use it in postmortems, keep referring back to the goal of an effort.
Three outcomes usually happen after a pattern is established and everyone’s noses have been rubbed in it:
This guy is a real one and knows the targets he’s aiming for and you’ll have an amazing time learning from someone like this.
This guy is like most of us and should be consulting his team for a comprehensive plan that gives the company a shot at achieving the outcomes you’re chasing.
This guy has no problem with failure as long as he’s the only one with his hands on the wheel. This is your exit music unless the arrangement is good enough for you wait out an inevitable failure.
kuncog@reddit (OP)
Thanks, good points. Hopefully I’m not in the 3rd category but only time will tell.
yetiflask@reddit
It's an actual startup. If you're worried about test coverage and tech debt you're doing it wrong.
kuncog@reddit (OP)
Well, I am worried because it’s a generated codebase that may or may not do what the CEO thinks it does but he is planning the next 6 months only along the happy path.
Tech debt in general is not a worry, working with AI slop is.
kernel_task@reddit
There's some nuance here. The successful startups I've been part of all have a messy code base that is almost instantly tech debt. They also get sold for millions. Part of it is inevitable: There's no way to not generate tech debt when you don't clearly know what you're building yet. Part of it is because you need to move fast or you're going to run out of money. You just need to make sure that the code base doesn't completely fall apart and is impossible to work in before you run out of money. At a younger start-up, that timespan is quite short, there's not a lot of code there yet, and so you can afford a lot of mess. You'll want to gradually start working on cleaning that up after you find product market fit.
However, pushing 5k lines of AI slop might make the codebase almost instantly unmaintainable. Bugs become impossible to fix if no human being in the whole chain understood what was going on in the first place, or at least the intention behind the code. There's a middle ground you have to find with your CEO but it seems like he's open to input.
kuncog@reddit (OP)
Good points and I generally agree about non-ideal code at this stage. If I can get to just regular tech debt I would be happy.
kingofthesqueal@reddit
Career founder at ~25 years old? I mean what career is there to have had by that point?
That’s definitely not the qualifications of a CEO or CTO.
Has he had a successful exit or 2 already? Gonna guess no since you said he also has a CS degree so I’m guessing he has less than 3 YOE total.
I mean if you guys have funding then I would bounce but this would be a giant red flag to me
SpaceGerbil@reddit
It means he gets to spend mom and dad's generational wealth by pissing it away on terrible start up ideas so he can write about them on his insta.
This startup is DOA
kuncog@reddit (OP)
You are close to the truth, but this one at least has some paying customers already, so there’s some optimism.
SpaceGerbil@reddit
Hey man, didn't mean to rain on your parade. It's just that I've been in this exact situation and it went... terribly. Mostly because you can't reason with a rich kid who thinks he knows everything and looks down on everyone. This is the same guy who told us NASA makes software without bugs, why can't you? This was in response for a request to hire QA
We had a saying those days : The beatings will continue until moral improves.
kuncog@reddit (OP)
You are not raining on any parade, I’m well aware of the negatives of my situation and you’ve pointed out the background of my founder accurately.
This is why I asked this question in the first place, I wanted other perspectives and anything is relevant to me that is similar to my situation.
kuncog@reddit (OP)
It definitely is a red flag and saw it when I joined, but the money is indeed good for now. That’s why I want to make the best of this situation.
30FootGimmePutt@reddit
I got so sick of him I quit in 9 months.
He was the most comically arrogant person I’ve ever met. Completely convinced of his own superiority and just an atrocious manager.
Nofanta@reddit
So he’s like 26? Good luck.
Adept_Carpet@reddit
The challenge is you can't code at a moving target. You have to find some way to get him to commit to units of work and see them through to completion, even if they are tiny units. Otherwise nothing will ever get done.
I also recommend trying to make sure the database schema is high quality. It might be good to hear from investors which metrics they want to see tracked.
Having a big messy JS file on the front end is one of those things you can chip away at, but a bad schema infects every part of the system and is a much bigger lift to fix. No testing is a shame because no one writes easily testable code if they aren't also writing tests but that ship has sailed.
kuncog@reddit (OP)
JS is also the backend actually, so imagine messy FE and messy Node
cleatusvandamme@reddit
Is this a company that will still be around in 5 years?
kuncog@reddit (OP)
Hard to tell, it exists for a couple of years already and we seem to have some product market fit so .. maybe?
justUseAnSvm@reddit
If you are late thirties, someone 15 years younger than you doesn’t have the experience to run a company, barring some outstanding outcome like building and selling a company in 2 years.
What this guy doesn’t see, and doesn’t have, is an actual plan to build something, so when people talk to him, that’s now the best idea, because he doesn’t have his own plan.
The single most important leader in a company is the CEO. If they can’t get the company to coherently move in a single direction, no one can.
I’ve worked in a bunch of startups: leadership problems in the good times means crisis during the bad times, when leadership is the most critical. Since I’ve already seen this play out, I’d cut my losses and leave.
kuncog@reddit (OP)
For now the money is worth it, but I’m also sensing this isn’t a very long term endeavour.
Material-Piece3613@reddit
Hes cooked