A new obstacle to landing a job after college: getting approved by AI
cnn.com
cnn.com
He's smart, motivated, very hard working, and has the necessary credentials on his CV, but when talking to him he's quite monotone and low energy. Because of this, he's failing these automated video interviews as he looks like he's not a "team player" or "personable".
While interviewing practice is becoming more aware of bias against things like this, the training data for these "AI" systems are going to be based on those interviews conducted in the past, often with questionable outcomes. The software is sold as helping "remove bias", but it instead pulls interview bias forwards in more extreme ways to the CV-review stage, preventing candidates from ever getting in front of a human.
While multiple humans on a hiring team may overcome some of their biases working together, a single piece of software trained to look for stereotypes will exacerbate those biases.
An automated system is the last-line in hiring; recommendations -- internal and external -- are the first-line. If one finds themselves in a position where they're manually submitting cold resumes, it's almost always more productive to start networking into the companies you want to work at, and getting the recommendation firsthand, turning your resume into a hot one.
> recommendations -- internal and external -- are the first-line
I'm afraid this is a very tech-centric view. Outside of a few specific industries or the very top levels, this is essentially unheard of.
My brother's educational background is biomedical sciences so he's looking for essentially lab work doing analysis for a hospital, drug company, or similar. There are a fair few jobs doing it, but they are relatively low level, have no "community", no real way to facilitate referrals.
In tech it's easy to "network into companies" because companies are so open with their hiring – they hold events, they sponsor conferences that are priced so that people can pay their own entry, and there are community events where you can meet people from them. This is very far from the norm, until you get to the golf clubs where you can mingle with other execs.
I’m only tangentially exposed to the biomedical field but I’m fairly sure this isn’t broadly true. I’m a member of a single cell RNA sequencing slack group whose members host a meetup twice a month. Most members work in wetlabs and many conduct research into the effects of various drugs on cancer.
The groups are probably harder to find (than programming meetups) but I doubt there are none of them.
I don't work in tech. But I've got about a 60% lifetime success rate (Job offers to applications). 100% once I got to the interview stage. And that's in a variety of industries: EMS, academic research, the energy industry, and civil/environmental engineering.
I swear nobody has any hustle anymore. I've never bothered to make a LinkedIn or go to "networking" or "hiring" events. They're a waste of time. If you're really out of your existing network (you're probably already doing something seriously wrong if that's the case), you'd be better off figuring out where you want to work and then waiting at a nearby lunch spot for an obvious group of employees to come in around lunchtime (or after work drinks) and start chatting them up. (I actually landed a job doing that.) Or better yet, find a CrossFit gym some of them go to. Sweat and bleed and bond with someone a bit before you leverage them as a recommendation. There's a million ways into an organization if you want it badly enough. If nothing else you can get super good intelligence on how to craft your application to be desireable.
Do your research, know your shit, know exactly what they're looking for before you ever turn in an application. Become that person to the core. Get any new certifications you need to be that person. Make every document you turn in to apply for the position fit that profile. Make every searchable piece of information about you on the internet align with that profile. Know the way they conduct interviews before you get there, and practice and rehearse the questions and flow of your responses in broad ways. Leverage your contacts in the organization to get information about each of the interviewers and how they think and approach interviews.
You know, hustle.
I'm going to go out on a limb and guess you're either a US citizen, or at least base this advice on the US.
I have only known 1, possibly 2 people who can make this sort of thing work here in the UK, people just don't do this.
Plus, CrossFit isn't really a thing here except in trendy bits of London. For many of these places there isn't a "lunch spot", people take their lunch in to their building in a business park where there's no lunch options or options for socialising.
Overall, while you don't work in tech, I think you're probably privileged enough to work in an industry that works pretty similarly. Most of your advice would be pretty good for me, but almost none of it would work for someone at the beginning of their career, aiming for a large company with out of town offices – a fairly typical starting point for many graduates.
I think the key for new graduates is to have been thinking about getting a job for the last 4-5 years. Don't start looking and preparing when you graduate, you're already behind.
I kept a job I started as a teen as a lifeguard for like 6 or 7 years even though there were much better opportunities available to me financially because I knew the stability was one of the best things I could bring to the table. Resume building.
But in addition to that, I started networking long before I left school. It's essential. Despite the fact that culture may differ in other countries, I don't think that fundamentally changes my advice. The tactics may differ, but the strategy is the same because human nature is the same.
I think it’s at the very least way too early to rely on this and at the worst just... terrible.
AST’s are already notoriously bad and this is just a whole new level of just plain bad. Time is precious, sure, but so are human relationships if time means anything.
It’s like nuance was completely disregarded in the pursuit of corporate contracts.
I personally think they _should_ handle it manually, and if the system breaks we should fix it in another way because this way sucks, but I'm not sure what that would be. More consultancies or temp agencies? Not sure that's a better option.
It just seems like we need to incentivize ourselves in the better direction all the incentives currently seem to lie in (at least what is coming to look like) the wrong places.
Assuming they don’t have the resources to give people a proper interview and vet a candidate, can you imagine the situation once those that do pass actually encounter on the job? I’d never work at a place that makes such kinds of decisions and it’s a good signal that the culture at the top of the company is completely broken.
Any company using one of these AI systems to combat attrition or justify growth is for sure a dumpster fire.
That’s just a couple of applications per job. How can the line managers possibly not have time to review just a couple of applications per person they’re apparently going to be leading?
1) Set minimum qualifications
2) For everyone who passes step one, run a lottery
3) That's it
Most interview processes are about as good as a coin flip at determining if a candidate is a good fit. People are biased (especially people who think they're objective, which describes a lotnof hring managers), and so are the systems built on our decision-making. So, apply similar logic as to the kind that supports index fund investment: cut "people" out of the decision-making, and make the process faster and cheaper by relying on a simple set of objective criteria and the way risk distributes across a population.
This is where things will go south for most big organizations anyway. They don't know what they need, or they can't specify what they need, and often this part is done by someone other than the team that actually needs people. So you find jobs where the requirements are 3 years of experience with C# and at least BSc in comp sci while the team needs someone to work on embedded C & Ada and they couldn't give a shit about BSc (or any other degree for that matter), just that you can do the job.
If they do, then would fixing the qualification issue be less expensive than interviewing et al.? Again, assuming that you have a choice of similar outcomes at different price points.
I've done interviewing for a couple of NHS organisations, from band 6 positions all the way to very senior leadership, and automated interviews were not used and would not have been used by these organisations.
> the NHS get something like 12k applications per week,
It's not really "The NHS". It's a bunch of different companies. They each have their own corporate structure and boards of directors. They have their own HR teams that advertise and recruit to positions.
There were (in 2017) 207 clinical commissioning groups, 135 acute non-specialist trusts (including 84 foundation trusts), 17 acute specialist trusts (including 16 foundation trusts), 54 mental health trusts (including 42 foundation trusts), 35 community providers (11 NHS trusts, 6 foundation trusts, 17 social enterprises and 1 limited company), 10 ambulance trusts (including 5 foundation trusts), and 7,454 GP practices.
The largest is probably Barts Health NHS Trust, which employs about 15,000 people. The smallest is probably Weston Area Health NHS Trust which I think employs about 1,700 people.
Imagine the horror if you actually started working for them! If their hiring is this bad, their working conditions must be even worse! So it's pretty kind of them to reveal their true nature in such an easy to identify way. ;-)
I read this as the software is sold as a fig leafs to wantobe rules who are pretty sure they have no cloths.
All you need to do is acquire the software, feed in a video clip of a male/female or white/black applicant saying the same things (maybe use deepfakes to ensure consistency in delivery), and if the score is one iota different, you've got the developer, at best, having to try to explain away a lot of ugly black-box logic, and at worst, a cut-and-dry discrimination case.
A human screener is probably somewhat biased too, but he's at least trained to provide a legally acceptable excuse for his choices.
i'm looking forward for the day when my AI "clone" - a very personable, cheerleader level bubbly and energetic deep-fake based of my face/voice - will charm away the dry pedantic HR AI on the other side of the screen.
It's a responsibility to help young graduates find an entry to a company. All it takes is that first entry.
My experience consulting was that the spectrum of work I could be asked to do was so broad I'd have trouble, despite having a handful of junior engineer friends, feeling confident I could hand off work to them; let alone a graduate with enough locality to be onsite. (and despite wanting to, as well, I'd love to give them those opportunities while selfishly allowing myself to scale further)
I'm not sure whether this is a problem due to lack of specialization on my part ("big data/distributed system design"), or lack of finding properly aligned graduates; I'd love your take on building those connections.
So I hire a Journalism or Business graduate, give them mentoring for the first month on-site, and being young and ambition they take off.
The principle thing in dealing with an Engineering organization (I tell them) is, when you don't understand what they're saying, say "I don't understand that. Can you elaborate?" Engineers generally love to explain their stuff. And it doesn't hurt to appear humble and eager either.
Anyway it works for what I've been doing (data migration, engineering data management, test fixture setup and deployment) since its more software configuration and vendor selection than it is technical Engineering. More a planning and coordination role. Especially since the departments I get injected into are already busy, and are glad to offload non-Engineering tasks.
It's a situation that's rich with opportunities to make yourself legitimately useful and productive. So the right candidate is not chosen so much by credentials, as skills in organizing, communicating and following up.
One very successful example had been a manager for an Americorp section, looking after 20+ young people as young as himself, and keeping them motiviated, supplied and supported. This person is now being groomed for a higher-level management position in the service company I introduced him into. I expect to see him as a Director or VP one day.
Btw, it doesn't hurt to have somebody at client companies that feels predisposed to look favorably on my services. Not only can I continue to get connected to whomever it takes to get a contract moving, I have a positive voice in their meeting rooms supporting us.
a) The supply of candidates greatly outweighs demand.
b) There is no clear way to assess the individual candidate's impact on the business' ability to make money, and the employees are essentially commoditized.
These environments have been known to be toxic for a while: filtering people based on bullshit 20-page personality questionnaires, managers hiring the most desperate ones who would tolerate the most humiliation from them, while every part of the process trying to a wink-wink nudge-nudge to others in order to land their nephew on a cushy position.
While AI definitely starts another round in the battle, I would advise anyone who's sick of this bullshit to focus on the economic output of their prospective position. If a company is looking for an AI engineer that can halve the size of their inference model without losing more than 1% of accuracy, they won't waste time scoring your story about your weak sides. They will look at your thesis and hire you if you already did something similar. The problem is that most of the jobs these days are about sitting quiet in a friendly team and smiling when the CEO shows off the team to the next round of potential investors. But those jobs won't give you much growth, and will end your career once you don't look young and cool anymore.
No, it does not work at all. It is fraud, and needs to be treated as a criminal enterprise.
I'm thinking the promises of this technology will never be realized.
Which aspect do you find farfetched?
The main goal was to filter people who have the biggest chance to go through the human interviews later in the process.
The only reason why it's called ,,AI'' now is because it's trendy. It can be just a simple linear model that takes education (University) and GPA score into account.
The simplest way to skip these filters is to get a reference from a person who already works at the company.
If you’re at the level of working for Google you’re presumably already at the absolute top of your field internationally and will know people at Google you can reach out to personally.
After the law suit my salary was increased 10% at Google, although I didn’t even work in US :)
This is exactly what happens. Most experienced engineers I know with a good resume get emails almost weekly from Google and Amazon in particular.
I’m not too worried about the long game though, humans are proving to be quite adaptable with tools like this one; an AI that optimizes your resume to match a given job description: https://www.jobscan.co/
However, in my experience, HR departments and hiring processes are anything but rational. HRs buy any kind of snake oil by the barrel, and hiring managers generally reach their position for reasons entirely unrelated to the ability to identify and hire above average employees.
It would be cool if there was a website where applicants could upload their videos and the hiring outcomes, to provide data from which organization's hiring biases could be inferred.
> The AI interview is too new, so job applicants don’t know what to prepare for and any preparations seem meaningless since the AI will read our faces if we make something up.
I strongly feel the market will work itself out: If AI interviews bias companies to hire hoop-jumping and actors, in half a decade, other, more traditional hiring companies will eat their lunch. If the AI interviews only deliver top notch candidates... it is working as intended, and who am I to complain about my own lack of prep or acting skills? Not like whiteboarding or being able to quickly think on your feet without access to StackOverflow is biased against me already, and companies who use these for hiring seem to be doing just fine.
https://www.reuters.com/article/us-southkorea-artificial-int...
"A new obstacle to landing a job: getting approved by AI"
Doesn't this same problem exist for anyone finding a job? It's probably even worse when you're looking and your 40 or 50 and even worse if you're changing careers. I'd bet AI drops anyone over the age of 40, almost certainly 50.
I'm not saying "these darn kids have it so easy today" I'm sure it's not, but my 1/2 educated guess is that the older you get the harder it gets to make it past these algorithms.
49 Right not to be subject to automated decision-making
..(1) A controller may not take a significant decision based solely on automated processing unless that decision is required or authorised by law.
..(2) A decision is a “significant decision” for the purpose of this section if, in relation to a data subject, it—
....(a)produces an adverse legal effect concerning the data subject, or
....(b)significantly affects the data subject.
I'd say that being rejected automatically for a job before being seen by a human would be a significant decision.
https://www.legislation.gov.uk/ukpga/2018/12/pdfs/ukpga_2018...
Probably something actionable under the nondiscrimination laws here, thankfully.
Seems like a constant gripe about talent shortages but then conflicting stories about how to filter an overwhelming supply of talent.
I'm honestly curious how both can be true. Who's lying, or is it simply unwillingness to train local talent?
> I'm honestly curious how both can be true.
There's an overwhelming number of applicants. That says nothing about their qualifications.
I don't know what Big Tech's numbers look like, but when I did hiring, the percentage of applicants that were even remotely qualified was usually in the single digits.
People explain away Google's high false negative rate, and desire to not do anything about it, by claiming they receive such large numbers of qualified applicants that it isn't worth caring about false negatives. How does that narrative jibe with the "talent shortage" or what you stated here?
https://elaineou.com/2017/03/10/technical-interviews-from-th...
> 2. A bill for 5 typewriter ribbons and 24 pencils totalled $9.85. If a typewriter ribbon costs $1.25, what is the price of a dozen pencils?
> The IBM Programmer Aptitude Test is from 1960 (? I think. It’s tough to read the date on the mimeograph). Is it just me or have technical interviews gotten much harder?
> I interviewed for a programmer job at Facebook several years ago, and one of the questions I got was, “Given a set of obstacle coordinates, write a function that finds the shortest path between two cells for a knight on a 3-D chessboard while avoiding the obstacles.”
> I remember this question because it was one of the few that I managed to solve. Probably a half-dozen others ended in fumbles. Not surprisingly, Facebook didn’t give me an offer.
https://elaineou.com/2017/08/26/the-mystery-of-the-vanishing...
> Some time ago, we compared technical interviews from the 1960s to those of today: They’ve gotten much, much harder. I suggested that the last few decades of lax immigration have invited overseas talent to raise the bar – it was unfounded speculation at the time, but some number crunching shows it might be true.
> According to this Joint Venture Silicon Valley report, 74% of Silicon Valley tech workers are foreign-born immigrants. A decade ago, 36% of Silicon Valley tech workers were born abroad. In 2000, only 29% were.
> Tech industry employment has increased from about 300,000 jobs in 2007 to 400,000 in 2016, so even though we created 100,000 engineering positions in the last decade, we’ve also displaced 88,000 domestic engineers.
I'm suddenly curious what other fields are staffed by 74% foreign-born immigrants.
Also, I don't know when CS degrees became common among programmers, but it didn't happen right after they were invented, I'm sure.
Not to mention, a vast number of people who could be good programmers wouldn't have had access to a computer before a first programming job, even if they had a college degree. Today it's much safer to assume that if someone is any good, they will have been exposed.
So I don't concede the point, anyway.
But, the examples above definitely do compare more and less computer focused things.
There is an overwhelming supply of people who claim to be talented but there is a dreadful shortage of those who actually possess it.
Also, I’ve noticed many companies are terrible at highering
tl;dr: the method works.
But government or big companies, well, its just bad news there, since they can't fail. They will just get more incompetent.
"Automated individual decision-making, including profiling" https://gdpr-info.eu/art-22-gdpr/