This one is short, but it’s been bothering me for a while.
A few months ago I stepped up as Head of Engineering, and alongside the function came the part of the job I’ve continually established the hardest: hiring. I’m fortunate to already activity alongside awesome people. But in a startup there’s a constant, quiescent force to keep looking for talent, since the instant you really need someone, you can’t oversee to delay three months for the ideal candidate.
Skills aren’t the entire story
I used to have expansive ideas concerning meritocracy. People have to be hired on their skills and nothing else. I couldn’t have been additional wrong. How you behave, how you connect to others, how you method problems and the energy you bring to a squad matter as much as your specialized ability, and frequently more.
From my admittedly anecdotal and narrow item of view, formerly you apparent a baseline of competence, your willingness to do the activity takes you much additional than knowing a bit more. In a startup this makes sense. It’s how small teams alongside constricted capability oversee to obtain on giants alongside thousands of employees, companies that can outbid you for all engineer on the market.
The blockers that have nothing to do alongside you
That’s why it hurts whenever fine engineers, the benevolent that are genuinely rare, can’t associate us for reasons that have nothing to do alongside them as engineers. They can’t get a visa. They live in another city. The agreement conditions don’t activity for them.
Lately, though, item alternatively has been on my mind. I accept using LLMs is now an essential part of application engineering. So in our interviews and specialized challenges, we anticipate candidates to use them.
Now put yourself on the another side
Imagine you’re the one looking for a job. You have no income. Job hunting today is mostly a numbers game, so you power be interviewing alongside five companies in a sole week. And all one of them hands you a difficulty anywhere you’re expected to use AI.
AI isn’t free. The tools that are really fine at coding need a monthly subscription or API credits. So you’re paying for them at what power be among the worst financial moments of your life.
The customary reply is: “It’s continually been akin this. You continually needed prosperity to get a job.” And that’s true. You need housing, apparel and food fair to be in a stance to apply. But none of those be lone to get you hired. LLM expenses are different. They’re ephemeral, and their lone intent is getting the job.
There’s no spotless answer
This puts me in an uncomfortable spot. I can obtain the position quo, knowing we’ll endure a non-zero figure of fine engineers merely since they couldn’t oversee a decent example for their challenge. Or I can try item different.
I’ve idea concerning this a lot, and there’s no satisfying solution. Giving candidates a business document for the difficulty is off the table: it’s a liability, and I can’t justify how it volition be used. Reimbursing what they expend doesn’t activity either. It’s difficult to validate as a endeavor expense, and it quickly drifts into a financial gray area.
What I can really change
If I can’t eliminate the cost, the one item I do authority is the process. So the goal becomes making candidates pay as small as possible, and lone whenever it really matters.
Move the difficulty later. Nobody should scorch tokens for a business that hasn’t had a appropriate conversation alongside them yet. If you attain our difficulty stage, it’s since we’re already earnestly interested, and the prosperity you expend is far additional apt to rotate into an offer.
Make it smaller. A difficulty that needs a top-tier example operating for hours isn’t evaluation the candidate, it’s evaluation their wallet. I’d fairly provision a tight, well-scoped issue that plant fine on a liberated tier or a essential plan, afterward expend the remainder of the period talking concerning how you approached it. That conversation is anywhere I study the most anyway.
Do it together. This is the one I akin most. Instead of a take-home, a live meeting anywhere the applicant drives and the AI runs on our side, on our device and our account. No liability, nothing to reimburse, no gray areas. And I get to see what I really attention about: how you think, how you communicate, and how you react whenever the example confidently hands you item wrong.
Weigh the code less. Honestly, the biggest alter is how much the difficulty counts at all. These days I provision far additional importance to a appropriate conversation concerning architecture and how you akin to build software: how you’d construction a system, anywhere you’d diagram its boundaries, how you think concerning testing, reviews and shipping. When an LLM can compose most of the code, the code tells me small concerning you. How you logic concerning trade-offs tells me a lot more, and that conversation expenses you nothing.
None of these are perfect. Live sessions are stressful for several people, and a smaller difficulty shows less. But I’d fairly choose a trade-off on intent than feign the issue isn’t there.
If you’re hiring too, ask yourself how much your procedure expenses person who has no earnings correct now. And how many fine group you power be losing without always knowing it.
Small footnote. If you live in Berlin and you are comfortably coming to our agency in Rosenthaler Platz and you affection startup environments, fire me an email to gerard (at) konvoai (dot) com. As a conversation starter, inform me what you think concerning RAG, how you've used it in the former and if there power be item improved for agentic AI assistants. :)
