Ask HN: Evaluate a programmer’s ability as a non-developer | Hacker News Reader

Ask HN: Evaluate a programmer’s ability as a non-developer

13 pointsRyanShook17 comments

If you’re good at sales and marketing and looking for a technical cofounder or first hire, how can you evaluate the technical person’s skill without deep knowledge of the underlying technology/languages? Are there any tests/services out there to help?

View on HN

First, accept that you likely can't evaluate that person's skill. Hell, highly skilled people often have trouble evaluating someone else's skill until they've worked together for at least several weeks or months.

Once you've got that idea out of your head, it's time to figure out ways to get shit done while minimizing the risk to yourself/your company. One great way is to start off with contracts, see how things work out and re-evaluate from there. If you have the budget, do a phase two that will require the candidate to hire/supervise another developer to get a task completed.

I'd avoid asking developer friends for help as hurt feelings will tend to spring up out of success or failure. I've learned that when you're starting a company, one good friend is worth one hundred good employees.

Great advice, thank you.

If you need any help or a sympathetic ear, I used to be obsessed with startups, but then I had a kid. My email is on my profile - feel free to use it as you need!

The best way is to have a developer friend/colleague evaluate the candidate's skill on your behalf. Preferably someone who manages other devs or has had to hire devs previously. There is really no practical way you yourself can make the evaluation.

How I imagine this going is that you have a developer friend who supports what you are doing but who isn’t at a place in her life where she can quit her job to focus on an early-stage startup. So y’all make an arrangement where you pay her to do interviews.

How should you structure the payment?

Sounds about right. If the founder was a good friend of mine, I would just screen devs for free. If it was a friend-of-a-friend type situation, I would bill my normal hourly rate (I am a contractor).

This is the only practical solution. You can get a feel for soft-skills/likeability but when it comes to hard-skills, the phrase "it takes one to know one" applies.

I would want to know 1) does this person have some basic competence at building something similar to what we need and 2) more importantly, have they shown initiative, grittiness, and quickness in creating things from scratch before. (In the form of their own projects they have worked on.) Beyond basic competence, those second traits are far more important for a potential cofounder than technical rigor. And a lot of technical people do not have those skills and cannot operate successfully in a self-directed environment.

Maybe try asking a technical recruiter for help or if you know any other experienced programmers if they can help with recruitment process. Otherwise you'll simply have to take the persons word. There are technical challenges websites like HackerRank etc that can give you an idea but I'd take them with a grain of salt as a fresh graduate with no experience can ace those.

Software development for an original product is one of the highest-risk activities that you can do in business. (It's like buying a bar without permits and expecting to open in a month.)

The best way to reduce that risk is to ask business owners you know, "Who wrote your prototype?"

You need 2 things:

1) Somebody who has, and can, deliver a working product.

2) Somebody you can communicate with when differences arise.

If it’s a first hire should I be looking for someone really good at the specific stack/language or a generalist who can solve problems and connect with the right talent?

If your company focused on building a specific technology that you will then apply to lots of industries, or on taking an existing technology and applying it to a specific business problem?

If the former, then its unlikely you are asking this question since you should be talking with the grad student who is already your cofounder.

So I’m going to assume you are building a startup like zipcar, taking mostly-known technologies and applying them to solve a business problem. In this case, you want someone who knows their particular stack very well (can debug their own package management issues, can choose good libraries, can navigate the docs quickly). You don’t need someone who has contributed to developing the language itself.

> and connect with the right talent

I’m not sure there is a large category of person who simultaneously doesn’t know their tools well but is well-respected among their contacts in the way that would draw other engineers to want to work with them. Maybe this refers to an engineering manager who hasn’t really been doing much programming in the past 3 years?

EDIT: actually yes, someone with engineering management experience who is a bit rusty with programming to any particular stack is probably a good choice—-as long as they previously have been efficient they can pick it back up again.

Pay them minimum wage and then give raises until the value they deliver to the business is met. Modern engineering strives to make programmers cogs via automation, doc and testing, so hire another random dev to work on your systems sometimes to validate the work of the first.

How on earth does automation, docs, and testing turn programmers into cogs?

Automation is removing the repetative work that a cog would do.

Docs and testing allow you to take on work of greater scope by being more sure of yourself and what your coworkers have built.

What kind of devs do you expect to attract with a minimum wage? Some can create negative value.

I think this is sarcastic but also maybe a little bit true?

There is nothing true about it. Try hiring devs at min wage and report back what happens...

Reply on news.ycombinator.com