Ask HN: Tips for hiring? It has been difficult | Hacker News Reader

Ask HN: Tips for hiring? It has been difficult

28 pointsaprdm37 comments

Recently we had an opening in our organization, the amount of very well crafted CVs we received it's crazy, they are really good.

Which imposes a challenge in itself, since we have to filter 100s of CVs for a role somehow.

Then in the interviews, it's clear that the person doesn't know how to go deep on the topics.

The conversation feels very unnatural, and very "buzzword driven".

In fact I am convinced that some of them are straight out reading from some AI prompt that "prepared" them for the interview.

I know hiring has always been difficult for both sides, but probably with AI's help the signal:noise ratio seems way out of the whack.

What have people been doing ?

View on HN

    What have people been doing?

I work at a public HPC center and we cannot compete with industry so we cannot afford to hire anyone with experience nor can we afford to hire anyone with an amazing academic track record. So we hire juniors with 0 YOE and a sketchy CV and we give them 6 months where their main goal is to learn the craft. And it honestly works amazing well. I don't know why industry is so focused on hiring only seniors because there's lots of good talent out there that could be had for much cheaper.

When we interview we don't try to assess how much technical knowledge the candidate has, but rather, of the things which they have gained experience in (and they are allowed to dictate this), how much can they confidently relay to us. From this we can estimate how much talent the candidate has. We get some interviewees with seemingly strong CVs e.g. particle physics PhD who does Kaggle in their free time, but then they are not able to explain in detail anything they have done. We also get some interviewees who have seemingly mediocre CVs e.g. bad grades, didn't publish their thesis and bare github, but they turn out amazing.

I guess my point is, experience is not talent, and some good talent does not sell themselves well and for the rest you get inundated with mediocre talent who know how to sell themselves.

In principle I agree. I worked in a similar industry until recently and it was much easier. Also because the influx of CVs for an opening was much smaller, I could actually read them into detail and have conversations. Because they were 10s of applicants over a month instead of 1000s.

I do hire juniors and interns as well. And I have similar issues for those. On interns we had 1000s of people interested.

If you are willing to hire in Canada plz let me know. I work as a data engineer with a few years of experience, so I’m not exactly your ideal employee. But I do believe I can learn the ropes in a few months.

At the same time I do believe that your hiring method works.

Regarding being able to explain (This is a reply as much to OP as this, as they both bring up the concept): Are you sure you're not testing for memory and verbal skills vice technical ones?

I've built a lot of cool shit! For the projects I've worked recently, I can talk your ear off. For one I haven't touched in a few years? I will sound like I don't know what I'm talking about, unless I reviewed it recently, or the right memories were jogged!

My team has finally found a successful routine. Here is how we do it:

* Our corporate recruiters are picky and yet hyper aggressive about finding/reaching people that seem qualified. It’s how they found me. I am not sure what their process is for finding people.

* Next they send me the resumes that pass their filter to see if I am willing to interview them. I am generally willing to interview anyone.

* The interview with me is a filter. I see if the candidate is qualified based upon experience alone. I do not bother with technical questions unless I am trying to see if they lied in their resume. So many people lie on their resume. Instead I mostly dive into their problem solving ability, their interpretation of technology, and general communication capabilities. You only need three things to be good at software: writing skills, an understanding of the platform/syntax, and people skills.

* The next interview is more technical and determines if the candidate is hired.

As a background I work for a defense contractor. People tend to skew much older, more experienced, and more mature. The work tends to be more predictable and polished compared to my prior software career writing JavaScript. People are expected to exercise initiative and figure things out instead hoping something like React, jquery, or AI will save their job. These sorts of pretenders are the people I try to filter out. I know I wouldn’t go back to that world where most people are under qualified and hyper insecure. I suspect that’s why people in this line of work tend to be older and why people tend to seek this work out when they could make more elsewhere.

Job Seekers, eager to get a job somehow, add the keywords in their resume to get shortlisted, without realizing that they might not go past the first interview. They don't realize that it's a waste of their time and the employer's time.

In my experience of placing software engineers in tech startups, the only way to find the best talent is to talk to them. If we give them coding test, they get done with it somehow but when we talk to them in the interview, that's when we get the real sense of their skillset. I know it's time consuming, but it saves you a lot of time and money in the long run.

I can’t share any recommendations, but know that it’s roughly an equally difficult experience from the opposing side.

Also, if you provide feedback to candidates whom aren’t faking expertise, major kudos and respect to you. Feedback loops were relatively common when I started my career and they’ve become mostly nonexistent today.

I second you on that. Especially for the senior positions as they are looking for an exceptional candidate experience meaning they should be getting updates regarding the application and feedback upon rejection. When they don't receive such experience, they don't wanna do business with such companies anymore as they feel disrespected. But since I started working as a recruiter myself, I noticed that providing feedback manually is a lot of work and recruiter's can't manage to provide feedback to all the candidates except the generic feedback which is just a formality but don't help the candidates a lot. We figure out a solution for it later and noticed 95% candidate retention rate by providing specific feedback to all the rejected candidates. So yes feedback works for both sides.

- Put best 30% into a pile

- Arrange 10min screening call with 20% (randomly selected from this pile)

- Use the screening call to verify the individual can convincing speak to the experience on their CV, and that they are a real person from the location they claim to be from (fake remote employees are a huge problem these days)

- Take a random sample from those who pass the screening call to do a tech interview

- Use the interview to get a better sense of their experience, culture fit, and end by setting a tech challenge

- Follow up on those who complete the tech challenge and make them talk through a few decisions they made (pick a few things in the code to discuss before the interview).

- Those who produce good code, can explain why they made the decisions they did, have good experience on paper and have past all previous steps are generally going to be pretty decent employees

Cut corners at your own expense. A bad hire will cost more in time and money long-term. Don't penalise those who use AI. Penalise those who use AI then can't explain their code.

How do you get the best 30%? 30% of 1k is still 300

Why do you think your company needs only the “best”? Wouldn’t a mediocre candidate do fine? By definition most companies are composed of mediocre individuals, and that works fine.

A bit tired of seeing every company out there thinking they are google, and cannot afford nothing but “the best”.

Where did I say that ? I am replying to a post that mentioned it.

I definitely want to hire the best matches to the role I am offering tho, but practically will settle for the ones that seem good to work with and that can get the work done.

https://posthog.com/newsletter/43-lessons-about-hiring-for-s...

https://www.recurse.com/hire

I haven't used them, but read the testimonials:

> ...sent us two profiles of solid candidates, one of which we hired!

> ...other recruiters sent like 50 CVs and we couldn't find anyone, and then RC sends a single candidate and we hired him.

How they do it: the Recurse Center puts more thought into collecting a small set of excellent candidates, then puts even more thought into matching them with employers.

The best ROI is personal referrals. The second-best ROI is interns->FTE or just new-grad from good schools. Industry hiring is hard, obviously top-tier talent is out there but finding it is exhausting. You have to actively look and participate and not just leave it to recruiters and automated filters.

Just pick someone at random. I feel like the outcome from most selection processes aren't any better than this. At least telling people it's a lottery is more transparent.

Tech companies usually require from you to:

- go through at least 4 stages (screening, tech challenge, systems design, chat with managers). You need at least 3 weeks to go through this in a sane manner

- You need to know about tons of topics (distributed systems, concurrency, microservices, k8s, ci/cd, databases, OS, networking). It’s common for engineers to know a little bit about everything, but when asked about something in particular in an interview, you cannot expect deep knowledge

- good luck trying to convince them that it doesn’t matter if you don’t have professional experience in $language, that you are rather an agnostic software engineer with experience in different progr languages and frameworks

- if you don’t craft your cv with proper buzzwords, you won’t get a chance to get interviewed. The ones who read your cv first are: machines (that are inflexible) and hr recruiters that know nothing about tech besides buzzwords

- good luck making a single mistake in the tech interview. Since the company has 100s of other candidates, they are only looking for the best of the best

Honestly, I don’t know whats wrong with using ai to “prepare” for interviews. I wish things were simpler, though (just come as you are, we talk for a couple of hours and if then you get a yes/no answer)

Networking.

Recruiters. Any recruiter who sends you someone like that, lose them. Find a recruiter that can actually judge talent - not perfectly, but enough to not fall for someone reading from an AI.

Yeah I agree, but I feel then we become very bias towards referrals and people who recruiters already trust - many developers despise recruiters for whatever reason.

I haven't seen a better approach tho.

As in everything, there are good recruiters and bad ones. I know several good ones who've gotten me good interviews - perhaps ask anyone on your team if they know any and consider using them.

Would putting more time into refining the job listing help?

What specifically do you want? Why? Why will this particular hire help the organization accomplish its goals? What values does a successful employee need to have? Why would someone who fits the ideal candidate checklist want to work with you?

It’s a good question but I don’t think so, short of requiring a cover letter with the application but I bet I would face a similar challenge there.

The candidates and CVs do look like what I’m looking for, until I interview them …

Are you putting out job posts or recruiting and trying to poach people?

I highly recommend poaching over advertising jobs and trying to sort through applicants.

we have to filter 100s of CVs for a role somehow

Using an AI, arbitrarily select 10 CV’s and move those candidates forward.

Discard the rest.

You don’t want to hire unlucky people anyway.

Good luck.

You can just shuffle and pick.

What's the point of AI here?

AI is still probabilistic. It just adds to your randomness.

Meh, just use a random number generator or dice. The "AI" is either just an inefficient frontend for an RNG, or it's using criteria that you don't know about and/or don't fully understand.

“Using AI” is a socially acceptable behavior in business environments.

And obviously, so is a calling a PRN, “AI.”

Did someone mention dice? There's a startup for that: https://chalculator.com/up/?world=diceroll&voiceChat=true&q=... First roll doesn't count. I think I'll write a blog post...

First to me, hiring from a job posting's crop of CVs is going at it the hard way: of course you get 100s of responses. If you have alternatives available: network, recruiter, user groups, trade show, anything really, then use that at least in addition to random CVs (which then are backup).

Second, I would suggest rounding up a few colleagues who can spend 20 seconds per CV. If 100s of these CVs all look basically the same - roughly impossible to judge a fit - then all these have failed already. Hopefully you are left with more interesting looking CVs which appear to actually say something (beyond keyword slop). A resume has one job. It shouldn't take deep reading to set aside 90% of them.

Which leads us to resume intake software. I understand why HR depts use resume intake software - they don't have the competence to read them. But normally the technical team does have the competence - and should be able to tell quickly keyword slop from not keyword slop. (AI will mess that up - but I don't think we are there yet.)

Some of the common resume formats out there fail us all 100%. They are not useful anymore. I appreciate candidates who notice what the resume is there for.

If you work is open-source you could try bounty to hire.

What are you hiring for exactly?

TLDR: Ci/cd for services on CSPs

A hiring manager once confessed to me that it was all luck.

> the amount of very well crafted CVs we received it's crazy, they are really good

please explain

It used to be that just quickly parsing through CV I could separate candidates that are worth interviewing.

Nowadays there seems to be way more CVs that are well written and crafted for the role. Some of the initial screening I did was clear the person didn’t have the experience in their CV

> without realizing that they might not go past the first interview. They don't realize that it's a waste of their time and the employer's time.

Still, not getting past the first interview is better than not getting there on the first place.

During my last job search, I was getting desperate and at one point ran my resume again some ATS scanning tools. I was getting dinged for not having verbatim buzzword slop in my resume (that matches the JD). Shit like "business solutions".

Reply on news.ycombinator.com