How we partner with software houses to find great developers

I prefer to have a team working together in the same office. But a great dev in another town is still a great dev; consequently, I'm building a remote team. Gemini claims we are setting “an unapologetically high bar”. Is it really that high? You decide. This post outlines our hiring process when collaborating with software houses and agencies.

Today, Sprend operates as a decentralised, remote team. At the moment, we are working with incredible web design and copywriting specialists from Croatia to redesign our brand and marketing pages. In parallel, it is time to bring in more developers. As we grow, we expect our agency partners to do quite a bit of heavy lifting before presenting engineers to us.

The checklist

At Sprend, we respect craftsmanship. We expect our software house partner to match us with a programmer while minimising interruptions to her coding focus time. Try doing it like this:

Phase 1: Cultural vetting

• DNA Evidence: Speak to three former colleagues using our 11-point Professional DNA framework (detailed below).

Phase 2: The two-stage pitch

Did phase 1 prove constructive? If so, the next steps are:

• Pitch the dev to me: The agency’s technical lead calls me on Signal (+46 70 714 52 99) to pitch the candidate based on the DNA evidence. Convince me why she is a perfect match.
• Pitch the project to the dev: Only after I am convinced should you present the Sprend project to the developer. Get her genuinely excited.

Phase 3: Selection

Now that mutual interest is confirmed, we request the developer’s time.

• Technical proof: If the developer is excited about the project, now is the time for your agency’s technical lead to verify her craft. You must watch the candidate step through code line-by-line (Asynchronous JS is our benchmark, see below).
• Chemistry check: If she passes the technical verification, we arrange a video call between the developer and me. This is for vibe and vision alignment—not a technical interrogation.
• The Sprend Challenge: If we "click" on the video call, we move to the ultimate question: Do we actually enjoy working together?

Our 11-point Professional DNA

Technical skills are just the baseline. We look for people who are ambitious and deeply care about their work (to get a feel for our internal culture, read about being a developer at Sprend).

Before submitting a profile, the agency must interview three people who have worked alongside the developer. If the dev doesn’t have extensive experience, the agency could talk to university professors and classmates, read her blog posts, and check out GitHub repos.

As an agent, before the interviews, make sure you understand each of the following eleven points. It helps a lot if you have experience working in a software development team. Don’t settle for yes or no answers; ask for specific stories.

1. She owns the task: She has seen too many team members abandon a task before completion. It could be someone who is “only the backend dev” and won’t touch the frontend code. It could be a project manager who lets another project’s problems dictate our success. The number one takeaway lesson from all writing on Agile methodology is to finish the current task before starting the next.
2. Deadlines hurt customers: A deadline is a very blunt tool, often based on the idea that pressuring people yields better results. It doesn’t. The result is a lower-quality product in every aspect, ultimately making the customer less productive. She knows this and will fight for quality.
3. Readable code is the foundation for great UX: She is naturally detail-oriented. She cares about code readability and solid architecture throughout the tech stack and beyond, into the user experience. She builds the app she wants to use herself.
4. Refactoring is not an afterthought: It’s tempting: let’s just do some “bypass surgery” here. But no, why would one increase the complexity of a system? Rather, the code she commits is better structured than how she found it. Refactoring is what she does as part of every task.
5. Digging deep: So she fixed the problem, but it was just too easy. That’s why she digs deeper. She makes sure she understands what is really going on and fixes the real problem.
6. Schrödinger’s code: She knows that untested code is exactly like Schrödinger’s cat: until you run the test and "open the box," your feature is simultaneously working and completely broken. Nobody needs to remind her to write tests.
7. Communicating freely: It’s a remote team. I would prefer it if I could roll my chair over to her desk to instantly double our intelligence. Since we can't, everyone has to proactively communicate—before, during, and after a task is implemented. In our team we talk about ideas, problems, and great solutions.
8. Low ego, high standards: I’m going to make changes to her code. She is going to fix mine. Let’s welcome feedback. Let’s learn together.
9. Pragmatic judgment: Every new tool or framework comes with a cost. As one thing becomes easier to achieve, another one becomes harder. She is pragmatic enough to choose the “boring,” stable tech over a trendy, hyped-up one because it is the smartest, most reliable choice for the product's long-term health.‍
10. Technology is but a tool (the goal is productivity and joy): We build software to help people climb the Maslow's Pyramid: we aim for our users to feel more confident, connected, creative, proud, and respected.‍
11. Attitude and humour: She loves software development. So do we. Our attitude is to have fun, let’s have a laugh. Let’s work magic together.

Technical proof: Running the code

At Sprend, trial-and-error coding isn't enough. We need developers who understand the deep mechanics of the language they are using. Therefore, before presenting a candidate, the agency must perform a rigorous technical screen. We expect the agency’s technical lead to run a practical test—for example, fixing a bug in asynchronous JavaScript code—and verify that the candidate can manually “run” her code. The candidate must be able to step through the logic line by line, explaining exactly how the execution works, what values variables hold, and how errors are caught and handled. The ability to read code has always been important; in an age of AI assistants, it is even more important.

Thoughts

This post is a work in progress. I’m trying to learn how to recruit great developers. Please give me feedback!

Read about our tech stack and our future plans for the product.