Re:Ask HN: How to make job search better for both the applic...
by jlokier on Thursday January 02, 2020 @10:25AM
(#21936084)
Some criteria I'd love to be able to search on (as a candidate/freelancer/consultant) are:
- Are they looking for someone urgently?
- Contract duration (if contract)?
- Fixed hours, slightly flexible, or extremely flexible "just get the project done" B2B style?
- Travel time if on site (by train - I don't want to drive).
- Quick start or lengthy interview process?
- Remote, some remote, or not at all?
- The compensation, or range.
- If equity is involved, do the math for me about how much it's really worth, by some standardised criteria.
- Would they be interested part-time to get particular expertise, even if they are looking for someone full-time ideally?
- Do they care about regular hours like most jobs, or would they be happy with a "this is the thing that needs doing, you figure out how" self-managed style?
- What is the field of work (not the tech stack). For example, I may know it's a (say) cybersecurity job, but I'd like to rule out some fields from (say) security of medical devices, social science research, food distribution, gambling sites, adtech, warships, etc. Ideally I could search by field or broader categories (even "civilian" would be a start) so I don't waste my time on things I know I'm not going to want to work on.
I try to search for thngs like that using clever keyword combinations, but it's not very effective.
LinkedIn's job filter refinement is quite good for this, but it doesn't cover all the categories I'd use. In LI, like with Ebay, given a current search resultset, it shows multiple-choice lists of properties that you can select a subset of to refine the search.
I wonder if something like OkCupid's old matching system would be good between companies and candidates for some kinds of work. In that system, the questions not only ask your answer from a multiple choice, they ask what your ideal match would answer, and how important each answer is, and the set of questions evolved, using a kind of statistical engine to deduce which questions were most effective for discriminating good and bad matches.
Things like geo-fencing, remote, freelance etc. would then not need to be pre-programmed specially, since they would tend to emerge via the statistical engine as high quality discriminators anyway. The most interesting bit is other high quality discriminators we haven't thought of may emerge as well (that happened on OkCupid, which is why the site became popular for "non mainstream" users).
To my mind the company-candidate matching process is mostly blocked on low quality of matching right now, which is why we have many candidates blasting out large numbers of queries or applications, at the same time as many companies blasting out job ads and yet both sides complain of a shortage of good matches and lots of wasted time.
- Are they looking for someone urgently?
- Contract duration (if contract)?
- Fixed hours, slightly flexible, or extremely flexible "just get the project done" B2B style?
- Travel time if on site (by train - I don't want to drive).
- Quick start or lengthy interview process?
- Remote, some remote, or not at all?
- The compensation, or range.
- If equity is involved, do the math for me about how much it's really worth, by some standardised criteria.
- Would they be interested part-time to get particular expertise, even if they are looking for someone full-time ideally?
- Do they care about regular hours like most jobs, or would they be happy with a "this is the thing that needs doing, you figure out how" self-managed style?
- What is the field of work (not the tech stack). For example, I may know it's a (say) cybersecurity job, but I'd like to rule out some fields from (say) security of medical devices, social science research, food distribution, gambling sites, adtech, warships, etc. Ideally I could search by field or broader categories (even "civilian" would be a start) so I don't waste my time on things I know I'm not going to want to work on.
I try to search for thngs like that using clever keyword combinations, but it's not very effective.
LinkedIn's job filter refinement is quite good for this, but it doesn't cover all the categories I'd use. In LI, like with Ebay, given a current search resultset, it shows multiple-choice lists of properties that you can select a subset of to refine the search.
I wonder if something like OkCupid's old matching system would be good between companies and candidates for some kinds of work. In that system, the questions not only ask your answer from a multiple choice, they ask what your ideal match would answer, and how important each answer is, and the set of questions evolved, using a kind of statistical engine to deduce which questions were most effective for discriminating good and bad matches.
Things like geo-fencing, remote, freelance etc. would then not need to be pre-programmed specially, since they would tend to emerge via the statistical engine as high quality discriminators anyway. The most interesting bit is other high quality discriminators we haven't thought of may emerge as well (that happened on OkCupid, which is why the site became popular for "non mainstream" users).
To my mind the company-candidate matching process is mostly blocked on low quality of matching right now, which is why we have many candidates blasting out large numbers of queries or applications, at the same time as many companies blasting out job ads and yet both sides complain of a shortage of good matches and lots of wasted time.