Job board where you only see openings which you are pre-selected/approved for the interview stage based on your qualifications

[roleyo5472]

This would be geared towards jobs that usually comes with a list of requirements that normally an applicant would have to manually go through one by one: years of experience, skills/technologies, education. e.g an Android developer with a bachelors degree, 2 years experience, knows Java Kotlin, bonus: .NET, c# etc

Problem:

The current process of job hunting can take days of browsing openings many of which contains requirements which you realize you do not meet half way through, wondering if you should apply anyways, deciding which to apply to and finally maybe getting a few interviews. I don't see an easy way to just filter out the search results that I lack the qualifications for, if there is a way, it's not obvious. If your skills become valuable enough, you'd be contacted by recruiters directly, and I imagine the process of finding the qualifying people on their end is just as tedious.

Value proposition:

The goal is to help employers/recruiters easily find the perfect candidates, and to help anyone receive the best job offers with zero time spent reading requirements. Success would be measure by the reduction in the amount of time to get a job opening filled for both parties.

Solution:

So the proposed solution would be an app where as a job seeker, you would simply create your profile and leave it.

And as a recruiter, you would be given access to a massive country-wide (if you allow remote) resume depository which you can query based on your specific needs; these queries would be based on weighted attributes that can be linked with the minimum compensation required by the applicant: (e.g. college degrees counts for 2 points, PhDs an additional 3 points, must have 2 years experience, each year above counts for 1 point, proficiency with C# counts for 1 point, etc. return candidates with min comp below 70k that has n points ,80k for n+2 points, etc) The app would return a list of candidates that best fits your requirements, and you can look through the list of individual candidates with the most tedious process taken care off, and decide whom to offer an interview, and the applicant would receive these through their side of the app.

Challenges:

A job board is only useful if there are massive amounts of users. One possible solution is to just have a service that manages openings / job applications for individual companies, until there are enough customers to make an aggregated job board to attract job seekers as users.

The technical challenge is that the app would need to actually be able to handle all the possible qualifications that people may have, for example it can't omit candidates with a degree in biology because they don't have a degree in biological sciences. It also needs to understand related skillsets, e.g. kotlin is related to java, aspnet core is almost the same as aspnet. This graph needs to be produced somehow, hopefully in some self-maintaining automated way.

Would you use this, why would you not use this as a job seeker/recruiter?

 

[tanuki_elex]

@roleyo5472 tl;dr; Been there, done that. Sounds good, doesn't work.

I used to work with a company that builds a recruitment platform for 200 of the Fortune-500 companies as well as mid and small-scale companies - and I was part of the team that built software way more sophisticated than what you mentioned. We had 300+ engineers, with 50+ Machine Learning engineers spread across 4 teams, with one team dedicated to finding the skillsets. What you mentioned in the last paragraph is what was a full-time job for three ML/NLP engineers working there - out of 50. So there are way more complicated problems compared to what you might think you already have an answer for.

We didn't just give "2 points for Ph.D." etc but built an algorithm that looks as past hires to understand the hiring trends (essentially, what the hiring managers are looking for) on a post by post basis - by algorithmically comparing job description, to the other related jobs, to the resumes and give a probability of both the candidate and position being fit for each other. We provide subtle cues to the candidates and a more elaborate explanation to the recruiters based on that.

Reality check:

Candidates said they don't see a benefit even after applying to the jobs that match the best

Recruiters said that they still have thousands for good equally similar candidates.

If you want to be paid, you need to focus on the recruiter, not the candidate. There are different software to maintain the interview processes/filtering - and different software that post in online channels - and a different one where candidates apply. If you can simplify the process for a recruiter by even 10%, they will be willing to pay you. If you wish to simplify the process for a candidate, you will probably increase the recruiter's work by 10%.

Another thing you might be totally forgetting about is - job requirements are not so simple.

Android developer with a bachelors degree => 100,000 candidates (either in your city or ready to relocate)

Plus 2 years experience => 50,000,

Also Knows Java Kotlin => 30,000

Might also know .NET, c# => 5000

(That's fully optional - why would you filter or rerank based on that?)

However, 5000 is still a too big number. If you can't pinpoint 10 candidates, your system is neither helping the candidate nor the hiring manager.

If you have a brilliant ML engineer who uses Python, has worked with PyTorch, you do not want to reject him for not knowing Tensorflow right? The fact that he doesn't know Tensorflow (your requirement) is meaningless if the other candidate is just an average ML engineer who uses Python and has worked with Tensorflow. Plus, once candidate understands what your system is doing, you can't stop them from gaming the system.

There are several other companies that are similar to my org. TBH, software that focuses on improving clients' experience didn't survive for too long. There is one big name that you might mention to contradict my point - have you ever applied through them? You send 100 applications and might get hardly one response.

EDIT:

In our meetings with our product directors and leadership of our clients, we realized that these "ease of applying" and "AI-based" and "nice fit" things were cool - but weren't the top criteria for them.

 

[roleyo5472]

@tanuki_elex Very interesting, thanks for the response. On the skillset issue, the easiest solution would just be to limit the market to tech jobs and then crowd source it as part of the sign up process, since we would require everything to be entered as part of profile creation process, we could have a prompt when entering into a field. e.g. "x does not currently exist within our database, please enter and alternative name or some related skills that an employer may look for" and as you type, existing #keywords are suggested and you can add multiple of these whether they're new or existing. If it does already exist, we'll suggest adding tags based on what previous users added for the same input. Limiting it to tech would allow this to be manually seeded. It's in the applicant's interest to not ignore it, and a small price to pay for the convenience of not having to browse for days. idk if ML is at a point where it gets more sophisticated than just scanning resumes and looking at the statistics of common occurrences anyways.

On the issue of the matching algorithm not seeming useful, you mentioned that your org provided an elaborate explanation of why the algo picked the candidates that it did, but what can the recruiter do if they don't like the rankings? I was thinking that we would just give full ultra-fine control to the human recruiter to adjust the mechanism as they wish and see the rankings change in real time: sort of like a split view with the controls on one side and rankings on the other.

By control I mean a bunch of sliders that you adjust/add/remove, each representing a weighted parameter. There are only a few categories of feasibly measurable and important parameters that I can think of: Experience, Education, prestige of each past employer (measure by average comp, need to get this data), gpa, prestige of schools (probably readily available), relatedness of classes taken, degree, skills, certifications (crowd source this data on profile page), presence and length of employment gaps, residence status, willingness to relocate, etc.

If you care about something, being up a slider for it, move it up and down depending on how important it is to you, and watch the ranking change in real time. You decide the algo, the app implements it.

A couple of ways this can be made to save time for a recruiter:

Obv if everyone just says I want the best and see the actual best in the world, it wouldn't help with finding actual hires. What can be done is to have min compensation as a required field as part of every candidates profile (possibly variable based on conditions benefits etc). Some job posts include the comp, resumes typically do not include expected min comp, but it seems like something that would save recruiters time looking for people that fits their budget. So the candidate pool can be greatly reduced by entering a max comp and filtering based on each min comp.

Next, the app can feature an integrated automatic scheduling manager for interviews a la Calendly, Booksy, etc. So you don't need to call anyone to offer an interview, you don't need to juggle times, worry about making too few or too many invitations and checking you calendar.

So the recruitment process would look like this:

1. Use the app for initial screening and produce a list of candidates that seems ok.
2. Hit save, optionally edit it manually , move people up/down, remove some
3. Use the app calendar to schedule a portion of the week to conduct interviews (only need to do this once).
4. Go to your editable ranked list, hit a button, and an interview invite is auto sent to the top 10 or some number of candidates, if your calendar is full, the invites are auto removed if they haven't been rejected already, and reappear as interviews close without a hire occurring.
5. Keep one-clicking the auto invite button as you need more time slots filled until no longer required. Close the position with the candidates list attached to it. (These interview invites can be preceded by customizable questionnaires, to steal from Ideal)

On the question of overall viability, what made me think that this whole thing could be viable was the impression that access to a large pool of talent was in itself valuable, and this large pool could be created by the offer of extreme convenience to the job seeker. E.g weworkremotely, a job board that claims 3m visitors (I assume counting one time), charges $300 to make one post. So assuming that the process that I described is not a game changer for recruiters, but if it's also not totally unusable, would the product as a whole be valuable (>$0 for an annual subscription) due to a sizable talent pool if one can be gathered?

 

[sarzy21]

@roleyo5472 Oversimplified answer to your question:

- From a negotiation perspective, it can be considered that the candidate will have the upper hand if the company will ask for a meeting, which translates in more money, more benefits, etc.

- On a similar note, you, as a company, by posting a job offer, will attract only the interested candidates. Otherwise, the companies might loose a lot of time communicating with people who don't want to work there

Depending on how you further segment the market, you might consider something similar to a headhunting agency or even a recruitment agency to deliver only the right candidates for the companies.

All the best.

 

[jamiee]

@roleyo5472 I think haystack is pretty similar to your idea and it's been out a while, LinkedIn is also the new Facebook