Stop vetting engineers like it’s 2021 — the AI-native workforce has arrived

You are already lagging behind if you still light engineers as it is 2021. We live through what I think that the most transforming technological shift of our lives will be, even larger than the internet.

The AI revolution accelerates at a pace that most of us cannot even understand. It’s not a hype. It is a herkalibration of what it means to build, create and work. Founders who are now preparing will lead to what comes next. Those who do not do that will surpass themselves by AI-Native startups of five people who work with 10x the speed and precision.

So how do you rent developers in this acceleration era?

You don’t screen them how well they write code. You show them to see how well they orchestrate. Let me explain.

AI -Flowness is actually the new literacy

Every founder wants an ‘AI developer’. But that term can mean many things. Are you looking for someone to build large language models (LLMS) in Python? Or someone who is competent in the use of AI tools to stimulate the speed and to reduce bugs?

Most companies need the second. But they don’t always know how to ask for it. That is the reason why AI fluity, or how well a developer can navigate and use a wide range of AI tools, becomes just as critical as knowing a specific language or framework.

The tool continues to change. But the meta-skill of learning how to use new AI assistants, evaluates their output and including that in your workflow? That is the sustainable advantage.

What is an AI orchestrator, and why do you need one?

An AI orchestrator is the essential developer archetype of today. They do not manually write each rule of code, criticism, error detection and refactor AI-generated output. They understand when they have to delegate for machines and when they have to apply their judgment. And they know how to communicate with AI agents such as colleagues.

At the same time, while AI is fast, it is not always correct. And it certainly does not have the specific needs of your company. So the characteristics you want priorities when assuming are:

• Architecture – the possibility to zoom up and design systems at a high level.
• Critically -Evaluation of considerations, making good decisions and choosing the right aids for the job.
• Communication – This is the big one. How well can you explain to a robot? AI does not do heuristics. You don’t get what you need if you can’t articulate what you want.

Just as we did not stop teaching mathematics because calculators exist, we cannot leave fundamental programming skills only because AI writes code. We need developers who understand the architecture, know when to trust AI and know when they should intervene and repair what is broken.

4 ways to assess the AI competence of an engineer

In response to the proliferation of AI tools, my company has overhauled how we screen technical talent. The traditional process of technical interviews, algorithm challenges and language-specific coding tests simply no longer cuts it.

Instead, this is what to do:

• Simulate Real-World problem solution. Ask candidates to build a position or to debug a problem, but do not allow them to write code themselves. Requires instead that they use tools such as chatgpt or claude, share their screen all the time, so that you can observe how they deal with the AI.
  
• Rate promptly. You are not only looking for the correct answer. You want to see how candidates frame the problem, encourage the AI and refine and repeat on the export. This exercise is more about determining the clarity of a candidate of thought and communication about syntaxism.
  
• Verify authenticity. Yes, people will try to cheat by sharing screens with someone else, having someone closer or resort to deepfakes. That is why you want to insist on sharing full screen and setting their camera. Let developers know that you are not trying to attract a “gotcha”; You want to understand how they work with AI every day.
  
• Test judgment. It is easy to get a working code from AI. The more difficult skill is to know if it is a good code, fits in with the system architecture and the right solution is for the problem. During all these steps you want to see if they can delete the bar of critical thinking about simple copy-paste.

What you should take into account with the adoption of AI

My team assumed that senior developers would get more out of AI. But what we found surprised us. In a series of surveys, Junior developers reported high productivity gain of AI, but often missed the opinion to catch a poor output. Senior developers, on the other hand, were skeptical or careful, which led to lower profit in the short term.

So we have built training for every level of experience. For juniors it is about delaying them, so that they are helped where AI sends them wrong. For seniors it is about informing them about integrating AI without losing control. In both cases, the goal is to unlock real productivity without endangering the quality.

Accept that change creates opportunities

Yes, this transition to AI is scary. And yes, there will be turbulence. There will be jobs that fade and new ones rise. But those who learn to screen, train and build teams around AI-enabled talent will write the future.

If you still hire engineers for what they can do alone, you miss the point. Start hire them based on how well they work with machines.

The future is not ai versus people. It’s ai of People, and those who adjust the fastest will win.

Jacqueline Samira is the founder and CEO of Howdy.comDie Elite Software Engineering teams in Latin -America builds and manages.