Advanced AI Based Candidate Evaluation Platform | An Interview with the CEO of Resume Sieve

Take a look at what Michael Yinger, the CEO, and Co-Founder of Resume Sieve, speaks about his experience with Cabot Solutions!

Introduction

Welcome to Cabot conversations, where we speak about innovative digital solutions for various industry domains such as FHR and healthcare. Joining us today is Michael Yinger, CEO, and co-founder of Resume Sieve. Hi, Mike thanks for joining us and love to get started with you making an introduction about yourself and Resume Sieve. Sure, thanks Shibu. Glad to be here and appreciate the opportunity. I've been in the technology space at hr technology space for almost 20 years and as such the founders came to me and said look at this idea we've got this application we've been using in our private company. We just sold that really enhances the recruiting process. It takes time out of the recruiter's day-to-day work by simplifying how resumes are evaluated and ranked and we want to go to market. We want to move this into a sort of commercial application.

Resume Sieve - Use cases

Tell me a little bit about Resume Sieve, what problem does it solve and what is the use case. I'd love to hear more about it. So we've got a handful of use cases today. I'll talk about two, maybe three of them.
The first is the amount of time that it takes for somebody to evaluate a resume really extends the time to fill for positions for recruiting whether or not you're reviewing them electronically or you're reviewing them on paper. If you're doing a thorough job there's a measurable amount of time you know. I'll just give you an example. We talked to one firm. They were hiring a recruiter. This recruiter was going to be hiring traveling nurses. They got 700 applicants in 36 hours. They shut off the app and the application process they printed all the resumes, three people, four weeks to go through the resumes to get to the people they were going to interview. That's a hell of a time commitment. So that's one issue. Now you could also say that well today with the way things are going on in the world there aren't as many resumes right because there aren't as many people applying for jobs but there still are in some sectors but that's certainly true.
Another example of a use case that the sieve can solve is well you've been collecting resumes. You've talked to people. How do you go back and find those who meet your new requirements? We can do that for you. What if you want to look at internally, what kind of skills do you have internally, well you put all your own people in the sieve and you run it through the process and you can find people who are qualified. We're talking to a large engineering firm about that particular use case that you know they've got 13, 14000 people load them all up and then use the sieve to find the people they need for their internal projects. Because it's you know they just would you know and would you remember did a good job the last time. You know the complicated solutions so those are kind of problems that we're focused on solving and where the system sits right now it certainly can do that we've just recently had a capability of collaboration and so hiring managers can collaborate directly with the recruiter in the system. So just making it easier for an organization to take advantage of the initial capability which is really streamlining that evaluation and ranking process whether it's new resumes old resumes, internal, external that's really the focus. So we can act like a repository of resumes of skill sets we've had people approach us asking to build a skills database for instance. Right. You know they might have 100 employees and for a specific project they might need specific skills for the people who need those skill sets and maybe Resume sieve can be that. Absolutely. It is ready to it can do that today if you know it's as simple as you can search for one skill or you can search for up to ten. Whatever kind of requirements that you had it learn itself to all three of those use cases and it does it very quickly. I was doing a demo here this morning. I sieved 303 resumes in six seconds. Okay. Yeah.

Resume Sieve Demo

You speaking of demo do you have. Could you show us a few use cases quickly? I sure I can. Let me get to a screen and so I'm going to start at the beginning because there's a new feature we just added that I really like and so I'm going to share my screen now. One of the things that we recognize early on is when you log in you just you're sort of in there in the system and there wasn't any real good way for you to figure out what to do next. So this is a feature that was just added to the system that you can keep and you can see it every time you log in or you can get rid of it and it recognizes who you are there are three things that you do when you're in the sieve.

You create a job.

You import some resumes and then you sieve them.

And so this gives you really quick how to do those things. So when you first log in you see the jobs that you have created down the left-hand side is navigation any sieves that you've done and you can save a sieve which is a snapshot for a future reference. Any jobs that you have and then of course any resumes that you've loaded. Now I created this job just recently and we'll go take a quick look at it without having to recreate it. You need to have a title you got to have something to keep track of it. You can upload a job description, you can copy and paste a job description, and it will populate this form for you or you can manually populate this form which is what I have chosen to do. So this is a project manager you can have you know the job title that you're looking for and I've selected some skills that they have project management, excel, project management software implementation skills, I want them to have worked for five years, I want them to have a bachelor's degree. Let me take a step back. We have 95000 skills in the database. It's possible you could ask for a skill that doesn't exist. I was having trouble with my demo just recently that I was trying to find a skill that didn't exist. I finally had to resort to typing it incorrectly so I'll do that right now. I'll just demonstrate that I'm looking for an animal feeder but I'm going to spell feeder wrong. Well, the system can't find it. Okay, no options. So can't find what you're looking for. Tell us what it is. We'll research it and we'll come back to you either well you spelled it wrong or here's another alternative or let's get this added to the taxonomy as I say we have 95000 skills in the taxonomy so it's unusual to find something that doesn't exist and we want to be able to help the user who can't find something at the moment. The same thing is true for education we have about ten thousand degrees, global degree categories so you can find a very specific degree if you really want to in this case I've just asked for a very general just a bachelor's degree, certifications we have about 3000 different certifications again. Global titles you can't find what you want let us know and we'll add it. Then we have languages, you can select the location you can either be remote or you can be specific and so I'm going to just for grins I'm going to say New York and so this is going to say I want to know anybody within 50 miles of New York you can adjust that and then visa and immigration status. Down the left-hand side are all the sections you don't have to fill in all the sections you need job title and either job function or skill everything else is optional this just tells you what you've selected and also allows you to navigate to that section of the job description. Next, I'm going to yeah I've done this before. There's here we go. I said this. Oh yeah. I don't want this skill anymore I see the system taking care of me. I started to create a skill and then I took it out well okay I'm gonna get rid of it. There we go. Okay. So now I have four skills and it says wait a minute it doesn't equal a hundred you gotta have something so I can have all the skills be the same which is what they are. Or I can adjust them which is to say you know I'm not so worried about the bachelor's degree. I want that to be five percent but I'm really interested in project management. You have control of the criteria as long as it adds up to 100 everything's good. Chance to review everything and then we can save it and so now my customer service or my project manager has been updated.

The next part of the process would be to look at the resumes now I've actually loaded some resumes. I loaded some this morning. I had some yesterday. I had some back in the past. We've got multiple ways to bring resumes into the system. The first is you can simply find them on either your hard drive or your cloud storage. You can then just up like you're opening a file you just open them and then they load or you can grab a bunch of them and drag them and drop them here or you can grab a whole folder and drag it here and so I'm gonna say this is today is the 23rd of July just gonna give it a title so I know which one it is. And then you're not going to see this because I'm doing it off the screen but what I'm going to do is I'm going to go find a folder for the resumes and bring those into the system so bear with me for just a minute and sample resumes and so I'm just grabbing a folder and I'm dropping it in there. And now it's processing. Okay. So all those resumes are loaded into the system. Behind the scenes, they're being parsed. They're being broken down into their constituent parts so that we can sieve them. And so I clicked on refresh all of them have now been sieved. And so that's great so our resumes are done and I want to evaluate this group of resumes. So I click on the box and it says well I can go to my sieve or I could come over here and I could click on sieve. Okay. So I've asked for a custom range for a specific group of resumes it doesn't know what job I want so I'm going to say I want to do my project manager test and I'm going to say sieve now and just that click it evaluated 910 resumes. Nobody met all my requirements that make sense because these are it project managers. Maybe they don't have all the specific requirements. So let's take a look. Well, the first guy doesn't have a bachelor's degree. Okay. It's a pretty sparse resume and he lives in Honolulu that's about as far away from New York as you can get so that makes some sense. Although he has been a project manager he has done implementation he does excel and let's see look at that. Yeah, he's got communication skills he's got implementation, recognizing agile as an implementation capability. There's some really good metadata here at the top of the system to assist the recruiter in evaluating the resumes quickly. What job are they in now, where are they working, what's their total work experience, what's their average tenure, what's their shortest tenure, and then any education or certification or awards. These are things that the recruiter might want to pick off the resume. But it could be a little difficult just because of all the different forms of resumes. You know, I didn't have anybody that met my requirements. So I'm going to look at my entire database. Oh let's see, I clicked the wrong thing. I'm going to look at my entire database to see if I've got anybody in these 300 old resumes that I've loaded that has the skills that I want. And so now I'm going through about 320 resumes. I still don't have anybody. I bet it's because nobody lives in New york. Yeah, look at that. I don't have anybody living in New york. Okay well. Let's say I'm not so concerned. I'm willing to have this person be remote and so I'm going to change the required location. Oh, I don't know if it is equal to 100. So I'm going to adjust them. And now I got a hundred and then next and I'm gonna save that. Yeah, I changed my requirements so I'm gonna have to sieve again and so I'm gonna sieve all 308 resumes or 12 or whatever it is I've loaded now, and just like that it has resieved them now couple of different things that have showed up here. One, we have people who meet all requirements. So let's go look at Kameisha and the score is calculated. The resumes are ranked based on the score they get. It's a relative score that varies from job description to job description. In this case, Kameisha has gotten 264 points because she's got way more than 60 months of total work experience so she got a bit of a bonus. On the other hand, she's got project management. This is an older resume the last time it shows project management was in 2020 so she lost a point because that was a period of time ago and then she's got a project manager and then the one that's often tricky she's got project management professional. Sure enough. Look at that she's got her pmp. All just aimed at helping the recruiter move more quickly through the process. One of the things that we've just added that I don't have set up yet is the ability to share a sieve with a hiring manager. Certainly, you know there's a subscription that the recruiter would sign up for to use the system but if they want to bring their hiring manager into the process they'll now be able to do that with no extra cost and they'll be able to show this to the hiring manager to work with them. Iteratively the hiring manager can review the results and then make a decision around who it is that they want to interview based on the results that they have.

Mike, I saw that in one of the resumes the keyword agile was highlighted. Now you did not put that keyword somewhere. Is there some kind of intelligence in it that if you search for a project manager that it also knows to look for agile or scrum or some things like that? Yes that's it they did. That was a good catch, Shibu very good catch. The parsing process is aided by ai for two reasons. The first is that resumes are all different. Physically, visually right and so the system is continually learning how to read the resume. You know, they put the jobs first, they put the skills first and they base it chronologically on all kinds of different things right. Right and sometimes hiring managers might search for java but you know it might spit out j2e which is a server. That's exactly right. So the taxonomy is organized into an ontology so that if you put java you said your requirement was java and what somebody had on their resume was beans the system knows okay beans that are synonymous with java because it's learned what are the similar terms and so agile it picked out agile well agile is a project management methodology so it recognized that this person has project management capability because they have done agile. So yes the ai is learning that as it's reading resumes and comparing things as we provide feedback on things that are missing that's a key function that you know I said today their 95000 skills those skills are grouped into an ontology so there are any number of skills that if you said project management, project manager, managed projects whatever the case maybe it's going to show up there and the system is learning those relationships all the time.

The number on the left the 821. Yup. And so that is a custom proprietary weightage algorithm. Yes, we have a proprietary algorithm the total possible score is a thousand, and eighty percent of the score is allocated to requirements. Twenty percent of the score is allocated to the profile. And then the score is adjusted based on whether they had more or less than the requirement and so forth. It's a relative number and so, for example, Kameisha got an 821 on this particular job description and if we go in and we look it's because well she doesn't have project management experience so I said you know this is this would normally be this could be as much as 200 she only got 160. She has a bachelor's degree so she did get a portion of that she did get some for project management and some for total work experience but she didn't get the maximum override. There's sort of a maximum override that you could get. So she's got looks like five years experience roughly as a project manager so she got some extra points but not the full. That's why she got 821. But the number itself is not important except that we use it for the ranking process right so each resume is evaluated against the criteria. In that particular job description and so in another job Kameisha may have a higher or lower score but she still might be the highest, the highest evaluated person. It just depends on what the skill criteria is.

And it supports pdf, doc, what kind of formats can I upload? Yeah today it's all the major formats can be uploaded so in this case in this sample we've got pdfs and we've got documents. You know it can read text, you know the one thing that it can't do today although we're working on it is it can't read images. Sometimes resumes arrive as an image because they've been downloaded from something or somebody printed it to a pdf and so it becomes an image we are adding over time the ability to actually ocr an image. The system is pretty smart. It's also smart enough to recognize something that's not a resume. You know maybe you grabbed a folder and there was a job description. In there well it's going to ignore that document or maybe it was something else you know maybe somebody emailed you and they included an excel spreadsheet well it's not going to try and upload the actual spreadsheet it's going to say yeah that's not a resume. So we are watching for things that do not you know won't parse in the way that you expect them to. I kind of have a good understanding of what these screens do is there any other screens that you'd like you to show. No literally you know it's a very simple system. Right we've looked at resumes, we've looked at jobs the one thing that we haven't done and we'll do that is that I can save a sieve and so I can save the sieve and it saves a picture of it and I can give it a name which is project manager test and today is the 23rd of July and so I can save that.

And I won't be able to edit it because I'm saving it but that's great because I wanted to this is the final one and I want to save this and I want to show it to my hiring manager and so I can go in and Ii can see that it's there it is saved and I can call it up. And if I wanted to make changes it wouldn't change it. Interestingly, I could also archive my jobs. Let's say I don't want to work with this customer leader job anymore and I want to move it to my archive and so yes and so okay and so now I've got some jobs in my archive. So some basic saving features that we've got today and you know we continue to enhance those and so software. Good, thank you.

What are your future plans?

And what are some of your future plans? Well, you know what I can actually show you that bear with me for a minute oh yeah. I can show you our future road map very quickly. I just have to open it back. There we go.

Okay. So we talked a little bit about the collaboration capability that is now available. Literally as of this morning. We're in the process of piloting an applicant tracking system integration. Let's say you have all your resumes in an applicant tracking system and you want to move them into the sieve for evaluation. We're working with one of the applicant tracking systems today on a pilot basis with a customer to do that integration. We're looking at adding some candidate reached-out features. So the recruiter's in the sieve want to schedule something with the candidate we're talking with a company about partnering and being having that capability within the sieve. In the longer term, we're going to add multiple languages. We're always looking to add additional insight. I talked about that bar at the top that showed you know how long somebody had worked at their job and so forth. We're looking at adding some additional insight maybe scanning social media looking for things that might not be accurate on the resume, any sort of benchmarks for what you might expect to find for those kinds of skill sets, and then finally bringing partners into the system we are talking with a company who has a skills assessment program and they can bundle those into you know sort of a standard skills assessment and the recruiter could then send the candidate directly from the sieve to the skills assessment process. So these are the general areas that we're looking towards advancing over the next several months to a couple of years.

That's interesting. So that means I consider my resumes and I can contact them right through the system and get them to do maybe a quiz or some kind of test and that comes back into the system is that right? That's exactly right. The idea is, you know, certainly I won't pretend we don't have competition in the market in this sense that there are systems that can do this but they tend to be much more expensive and they tend to be targeted at much larger companies. Our focus is primarily on small to midsize companies today along with the partnerships that we're attempting to build and these are companies that typically don't have either the time or the money to learn a complicated system and we just want to put simple to use set of tools in the hands of those who are doing recruiting it could be a recruiter, it could be the office manager, it could be the business owner. I was talking to a business owner who happens to be my brother-in-law he runs an auto shop that also does window tinting and you know we're looking at using the sit for him and he only has 25 employees but he hires probably four or five people a month because of turnover you know it's an hourly job and he has to deal with turnover but he does it all himself. You know he pulls his job he pulls people off of Facebook or craigslist that's exactly the kind of use case we're talking about putting a simple to use tool in front of somebody and you save them all that time that they're spending going through the resumes manually.

How can users sign up for a demo?

And how can a potential user sign up? Is there a free test they can do? Yep, they. At our website and so let's go to just because I can we'll show you the website and so yep our website resumesieve.com we'll go back to the homepage. We have a 30-day trial. It's there everywhere. Feel free to use it we're also doing some crowdfunding and looking to get some participation from some people to work with us to get to sort of that next generation. And so this has a link to our crowdfunding campaign if you're interested uh the whole idea here is in addition to seeing some comments we've gotten from people who've been using the application. You know come in take a look and sign up there's no cost. It's 30 days we don't take your credit card. The idea is we'd like you to come in and try it talk to us are there features that you think should work differently should work you know that doesn't suit your business model we're really in the stage, where we're having those kinds of conversations with people, how can you best use the system, what kind of features are going to get you to where you need to get with your particular recruiting issues.

Your experience in hiring a development partner?

How was your experience in finding a development partner and do you have any advice for other startup enthusiasts who are trying to build a new product and come to the market?

Yeah so the first piece is you know might be advice that you don't take that don't follow as soon as I put my new job in LinkedIn, my inbox started to fill up with people who were sure that they could do the development that I wanted to have done. And I literally if I spent the time necessary to talk to all those people that would do my full-time job. So it's kind of an interesting thing to be aware of as soon as you put in you know that you're on a startup or that you're the CEO or whatever the case maybe you're going to get a lot of people reaching out to you. They're going to be from all over the world. Literally, there are tens of thousands of startup companies. As I talk to other startup leaders they're really two ways that startups get moving that you know it sort of fits into two buckets. The first is they do it themselves because they have either the skill sets or they got friends with the skill sets and so they build something themselves. May or may not be cheaper. You know we talked about doing that ourselves, you know hiring a couple of people and we decided that it should be a lot more expensive. The other way is to find a partner and I described earlier that's the route we chose. We reached out to people that we trust, we asked for recommendations, we talked to references and then we went through a process where we gave you know the folks that we were looking at a set of issues to resolve for us to come back with a potential solution. We probably talked to about a dozen firms. When we went through this we could have talked to a lot more but we were in a hurry as startups often are. So the decision process took us perhaps six to eight weeks from start to finish which is a lot of time at the beginning of a startup because you're not doing anything if you don't have the development team. At that point I did have a CTO but he didn't have anybody to direct. He didn't have anybody to work on things so that was good it was good for us he was able to work on his road map and get some of his stories written and so forth. But you know as you talk to these various firms you want to understand you know do they know what it's like to work with a startup how have they done that how do they move your tool forward, what's the project management capability, these are all the kind of things that we looked at and then overtime is certainly you know as our relationship with you at Cabot has evolved we've worked together to find a good operating methodology, right? I mean we're certainly operating under an agile program that there's no question about that. And you know where's the dividing line who does what work, how much of it is going over the other side, we made a conscious decision early on for example that we wanted to retain a substantial amount of the project management and the intellectual property development on our side rather than just saying to you here's what we want to do go develop it and come back when you're done we wanted to be involved in that process. One because we wanted to be involved in the process right we wanted to keep ownership of it but two we have a certain amount of industry expertise which then can supplement your technology expertise and so as you're talking to other firms. You may not find a firm that has an understanding of the kind of project or the kind of tool the kind of project that you want to run. They have all the technology skills that you want and so you need to work collaboratively with them and so that's been that was a learning experience for us and I think we've developed a nice smooth pattern as we go through it. For example, our product management meetings are not just us not just the people on the Resume sieve side the project manager from Cabot sits in with us to hear what's going on it's we're not asking her to do things because until we hand something over you know her job is to make sure that the work that's being done by the developers gets done and she gets to hear the conversations we're having where do we think we're going and by the way, she participates as well. So it's really helpful like you know if you're going to do that maybe we should do it in this order or you should do that you know along with this because they tie together because it's the same developer all those kinds of things that can help us more effectively utilize the resources we've got. But it's finding a trusted partner and then you know quite frankly you know we've been with Cabot now going on a year and you know it's developing a trusting relationship and working through it.

So it almost seems like it's a kind of an extension of your own team it's not just another it's not just hiring another company it's an extension. Yeah we again it's a conscious decision that we've made that you know there's you know should you bring it in inside or should you know have a trusted partner and we continue to operate in this idea that intellectual property and you know the oversight, yes but you know we're not big enough and you know a firm like Cabot you guys can do it more efficiently and more effectively. It just really works for us.

Awesome. I think that's all the time we have for today. Thank you very much, Mike. Hope you know good luck with the software and future endeavors.

Thank you, Shibu. I appreciate the time and I appreciate the interest and I appreciate all the support that we're getting from your team. So thanks and enjoy the rest of your day. Thank you. Have a great day. Okay.