Ask a founder what their new CTO will do on the first Monday and the answer is usually some version of “build the app”. That describes a lead engineer. The distance between those two jobs is where most first technical hires go wrong, and the bill for getting it wrong arrives about a year later.
Three different jobs share one title in most founders' heads. Pull them apart before you write the job description, because they cost different amounts and they solve different problems.
Three jobs, one word
The contractor
You define the work, they price it, they deliver it. This is a clean transaction and it is the right answer far more often than founders expect. If the thing you want is well understood and you can describe it, buy it. What a contractor will not do is tell you the thing is wrong. That is not what you hired them for, and a good one knows it.
The lead engineer
Owns how the product gets built. Chooses the tools, writes most of the early code, hires the next two engineers and reviews their work. Cares deeply about whether the software is correct. Does not usually decide what the company should be building, and does not expect to. This person is hireable at a market salary, and if it does not work out you can replace them without the company changing shape.
The CTO
Owns the relationship between the technology and the business. Build or buy. Which technical risks the company is accepting and whether they are worth accepting. This person sits in fundraising conversations and answers to a board. In a five-person company they also write code every day, which is exactly why the three roles blur together.
When you actually need one
Most pre-seed companies do not need a CTO. They need someone who can build, who has enough judgment to say when the plan is wrong, and who will still be useful when the plan changes. A CTO title on that person costs you equity and sets an expectation that gets awkward when you raise and someone more senior joins.
Here is a test. Name a technical decision from the last six months that would take more than three months to reverse. If you cannot name one, you are not yet making CTO-shaped decisions. If you can name three, you have been making them without anyone qualified in the room, and you should fix that before you make a fourth.
Testing judgment when you cannot read code
You are not qualified to evaluate whether someone can program, and you should stop trying. Get that from a paid trial or a reference. What you can evaluate, better than most engineers can, is judgment. Four questions get you most of the way.
Ask them to describe something they built that they would build differently now. Weak answers name a technology. They picked the wrong database, the wrong framework, the wrong cloud. Strong answers describe a decision about scope or sequence. They built the whole thing when half of it would have answered the question. That difference tells you whether the person thinks about outcomes or about tools.
Give them your current plan and ask what they would cut. Anyone who agrees with all of it in the first hour is selling. You want someone who reads a plan and immediately finds the part that is doing the least work.
Ask what would have to be true for this to fail technically. Vague answers are a bad sign. You want specifics about your product, not a general lecture about scale.
Then ask them to explain something technical from their last job. If you cannot follow it, that is real information. You will rely on this person to explain hard things to you and eventually to a room full of investors. Someone who cannot make you understand their own work will not do better with a board.
Why the equity conversation is different for this hire
The first technical hire sits somewhere between an employee and a founder, and the equity numbers reflect that confusion. A lead engineer joining a funded company, on a real salary, with a product already built, is usually somewhere between half a percent and two percent. Someone joining before any money, taking little or no pay, making the decisions described above, is a co-founder. That is a different order of magnitude, and calling them a “founding engineer” does not change what they are owed.
The failure is paying one for the other. Co-founder equity for employee commitment is the obvious version, and it hurts loudly. The quieter version is employee equity for co-founder work. That person stays fourteen months, leaves with a small vested stake and all the knowledge of how the system fits together, and you find out what it was worth when you try to replace them.
Four year vest, one year cliff, in writing, before anyone starts. If a candidate will not accept a cliff, they have told you how long they intend to stay.
The failure that quietly ends companies
You have a spec. You hire someone good to build it. They build it well, on time, and it works. Then customers tell you the product is aimed at the wrong problem, and the person you hired has no opinion about what to do next, because opinions were never part of the arrangement. You are now paying senior money for execution and you still have nobody to think with.
Hire someone to execute a decision you have already made, and you have quietly made that decision permanent.
The prevention is cheap. Bring the technical person into the decision before the spec is final, even two weeks before. Hand them what you have already decided and ask which part they would reverse. If they have nothing to say, you have learned something important while it still costs you nothing.
Practical steps
- Write down every technical decision you have already made and give it to them on day one. Ask which one they would undo.
- Run a paid two-week trial before you make an offer. Serious candidates say yes to this.
- Take a reference from someone who reported to them, not only from someone who managed them.
- Agree in writing what they decide alone, what you decide alone, and what needs both of you.
- Set the cliff at twelve months and hold to it.
None of this protects you from hiring the wrong person. It protects you from finding out eighteen months late, which is the part that actually costs the company.