If you are not technical, it's probably not a good idea to interview developers yourself. You can't tell a sound answer from a confident one, and a wrong hire comes out of your budget. Hire a fractional chief technology officer, a senior technical leader, or a very senior engineer. You only need them for a few hours a week. Let them run the interviews, review the code and mentor the team. Keep the decisions that belong to an owner: what the software must do, what it may cost and who owns it.
Why a founder's interview fails
Founders tend to pick the person who sounds most sure of themselves. Confidence says little about how the code will hold up.
A list of questions doesn't fix this. You can learn the questions, but you can't judge the answers. A fluent, wrong answer sounds the same as a fluent, right one.
Why the cheapest hire costs the most
Founders who want to save money tend to make one of three hires: a group of junior engineers overseas, a recent graduate with no job experience, or an overseas firm that claims decades of experience and charges far below the market rate.
The first two fail for lack of training. These developers have never worked on a team, have no senior to learn from, and have never kept a product running for years. Training teaches them to finish a product and not just a demo, to design for new features, and to write code another person can read.
The third hire catches founders off guard. Decades of experience at a fraction of the price looks like a bargain. But experience costs money, so a firm that charges far below the market must make up the gap somewhere. The next section explains how.
In all three cases the software works at first. Then each feature takes longer than the last, and only its author can read the code. The low hourly rate hides this, because the cost arrives later as extra hours.
Founders in this spot often decide they are bad at their job. They are not. The developers lacked training, or the firm had no reason to stay and fix the work, and nothing in the process caught it. This usually ends in a rewrite that more than doubles the cost of the project. Repair or rewrite shows two real cases.
Lower-cost developers can do good work once a senior person sets the structure and reviews what they build. The cost problem starts when nobody does. There are excellent developers in India, Chile, Mexico and the United States. Where someone lives isn't the problem. Training and review are.
Why a very low price means no one stays to fix it
A firm that wins clients on price must keep its work at the minimum the client will accept and its costs at the minimum that still earns a margin. With margins that thin, volume pays. Moving fast and moving on to the next client earns more than staying to get your product right. Most of the margin is made during the first build. In maintenance, billable hours drop. On a fixed-price job, the way to raise the margin is to build the cheapest product possible. If you want cheap, the engineers must cut something. Nobody is acting out of spite. The market produces this result.
Here is how it looks to a founder. The team delivers the first milestone, and it has problems. That is normal, whatever the size or price of the team. The first look at any work nearly always brings tweaks. Two things then go wrong. Founders rarely have training in quality assurance, the testing of software, so they can't test the application fully. And few keep pushing back on each tweak. Once you accept the milestone and the work moves on, the chance that a problem gets fixed starts to fall.
Platform ratings tell you less than a phone call with a past client.
Three steps protect you:
- Ask for a video call before you hire, and treat reluctance as an answer.
- Ask for referrals, not just to see past work. Phone those clients and ask what went wrong, instead of reading their ratings.
- Before a milestone begins, state in writing everything you want from it.
Be this explicit because no team wants to build something it wasn't asked to build. If you didn't ask for it before the milestone began, don't expect to find it in the finished work. Writing it down is how you get what you want.
Why AI doesn't replace a senior engineer
Some founders now skip the hire and build with an AI tool such as Claude. That can work for a simple first version. It fails when the product grows complex or security matters. The tool loses track of the whole system, and a founder who can't read the code can't see it happen.
A wave of rewrites of AI-built software is now reaching the market. Freelancers used to be paid to start projects. Now they are paid to clean them up. Some projects are in such a state that no one will touch them, and the only offer is a rewrite.
The same happens when a developer who can't design an application relies only on AI. The software throws random bugs. The AI gets confused by its own earlier work. It also adds bloat, because it needs more and more code to understand what it wrote. What AI tools change for a founder who owns software explains why structure and review are still worth paying for.
What it looks like when it fails
One founder recently asked for help with a project that grew from 500,000 lines of code to about 1.4 million in six months. That is nearly three times the code in half a year. At first, every release went live without failing. By the time they wrote, only about 40% did, so a fix or a new feature broke the system 60% of the time. They also described a loop of bugs they could not end. Each fix produced the next failure.
This ends in a rewrite. The code is so poor that none of it can be salvaged. The founder paid for 1.4 million lines and has nothing to keep.
The worst part is that paying customers use this system. Every week is a race to stop them cancelling their contracts. The only way to end the failures is to rewrite everything, which could take months to a year. Few things are worse than watching your customers suffer, with no room left for anyone to save you.
Size is the warning. A product that triples in six months has not tripled in value. More code means more to read, test and host. Without a senior engineer to set the structure and review each change, every addition makes the next one harder.
Hire a fractional CTO or a very senior engineer first
The proper fix is to hire that senior person first. They do three jobs. They run the interviews. They review every pull request, the step where a senior developer reads each change before it joins the product. And they mentor the developers who write the code, which is the training the cheap hire never had.
They cost more per hour than a junior. You only need them for a few hours a week, and that cost is lower than the rewrite that follows the cheap route. What it costs to hire a software developer gives the rates for freelancers, employees and agencies.
This person should have no tie to the candidates. They should not work for the agency that supplied the candidates or hope to win the work themselves. Have them review each candidate's code. Don't skip this step, because a candidate can't bluff through a code review. The review should cover how the code is built, how it handles errors and whether it is secure.[1] Many experienced engineers will sit in on final interviews or review a trial project for an hourly fee. A 30-minute review call after a trial can save weeks of a bad hire.[2]
Ask them to answer these about each finalist. Would you be comfortable inheriting this code? Has this person worked on a team with a senior developer to learn from? Is the structure open to new features? Could a cheaper developer maintain it after the author leaves? The last matters most if you want software that costs less to run.
If you want help finding and vetting that person, Moosara's technical recruitment service can do it.
If you already have software
Many founders reading this have a half-built product and want a second developer. The order changes. Have a senior engineer review the existing code first. The review shows what to hire: someone to repair it, finish it or replace part of it. Four checks on what you already paid for cover who owns the code, whether a stranger can run it and what it costs to host. Do them before you interview anyone. If you're already stuck, Software Rescue is how we review and fix what's there.
Where to start
Before you interview anyone, get a senior person involved. If you want to talk it through, book a free 15-minute call. We'll look at where you are, what you're building and who you should hire first.
Sources
- Lev Perlman, Metamindz, "How to Hire Your First Developer When You Can't Code" (published June 2026, updated September 2026) https://www.metamindz.co.uk/post/how-to-hire-your-first-developer-non-technical-founder-2026
- Chris Tottman and Ruben Dominguez, The Founders Corner, "How Non-Technical Founders Can Hire Great Developers" (July 23, 2025) https://www.the-founders-corner.com/p/how-non-technical-founders-can-hire
