#30: AI can help you get a job.

This morning we're going to help you get a job. Who are you?

Yeah, I'm Eric and the founder of jobrite AI Job right is basically your personal job search AI job search copilot that will guide and ultimately entire job search process. Because the job search is so efficient and time consuming, we want basically want to bring you a personal recruiter which can save you 80% of time and also bring you more interviews. That's awesome.

I need more interviews. Tell me a little bit about what your system does. Because we have LinkedIn, we have our resume, we have email. Why do we need you?

Yeah, for sure. I think that's definitely the first question you ask us. With LinkedIn dominating the market with so many jobs, why do I still need job right? I think yeah, we start with the pain points especially so right now if on LinkedIn. Right. So even there's a. You still need to rely on yourself to do the job search right? Find the jobs you think it's relevant to your skills and you are qualified. You need to do the research about the companies and then you need to look at the job description and customize your resumes and even apply for it. It's really time consuming. And also we found out actually majority of job seekers, they're very talented in their space but they actually don't have know how to do job search correctly in this they might miss out, miss out a lot of great opportunities. They don't know maybe how to write a good resumes even they don't know they use LinkedIn for networking. But not many people know how to do Networking correctly on LinkedIn. So our answer is pretty simple. You still use LinkedIn but job rights is basically a tool to help you to automate all of that. It's like giving you per recruiter with top expertise and on your side and helping you to do the work 24 7. Yeah, yeah.

There's some discussion that LinkedIn is the terms of service on LinkedIn keeps you from using automated tools like this. Do you have a position on that or is it okay to use. Am I going to get shut up if I start using your tool and start doing a lot of automatic stuff?

No. So job Reddit. I can definitely do a demo. It's an independent setup. We source jobs not only from LinkedIn, LinkedIn indeed the recruiters. We also actually source jobs from more than 200,000 company site like Apple X. They never posted jobs on LinkedIn. So yeah, it's totally legal because we use all the, we use all the public available information and yeah, so our tool stay perfectly with LinkedIn. So it's not a plugin, it's a dedicated website that you can use without any problems. Yeah, very cool. Help you to identify the jobs on LinkedIn. So when we identify it, when you click it, you go to LinkedIn and you do the work on the LinkedIn part.

Yeah, it can take my LinkedIn and my resume and help make it better. Particularly for the AI filtering that a lot of companies use now. Right, right. A lot of companies use AIs to read, you know, like at Microsoft, I personally got 30,000 resumes sent in with a video. You know, it's like, whoa, that's a lot of resumes to go through. Right. And now these companies are putting those right into AI and the AI is telling them whether to contact this person for a job interview or whatnot. A little sense of what it does for making your resume better.

Yeah, for sure.

And I'd love to see a demo after that.

Yeah, for sure. Yeah. I think nowadays especially so, whenever there's a new job position, a job posting, there's like easily more than 200 applicants out there. And also the company side, they usually use ATS to score your resume and to analyze your resumes. So that's why I think it's super important to make sure your resume is first, ATS friendly, it can be parsed correctly by ets. Second, you definitely, because there's so many applicants, you definitely want to make sure when the recruiter look at your resume, your experience is highly relevant. Right. So especially like if you have a work experience with like many years. Right. So you have a lot of experience, but how you can tailor it and to be more relevant to align with the requirements of the JD is very important. But to be honest, because like as a job seeker, you maybe apply 100 jobs per week or something like that, but it's almost impossible for you to tailor resume for each job. Even they take 10 minutes. Think about that. It's like 1000 minutes too time consuming with job. Right. We want to help you to do it for every jobs, maybe within a minute or. Yeah, so that's basically the goal.

Very cool. Could we get a demo and then after the demo we might talk a little bit about how it works underneath the hood.

Yeah. This is our website. Currently you can either use this on PC or we have mobile version as well. But definitely the PC version has more features like resumes and it's more suitable for PC. Yeah, I will just quickly go through that first. I think right now JobRight has already within six months we have more than 100,000 users using us like Daily Base. And the number one reasons first they come to our site is because at jobright they can find the most fresh and relevant jobs across all the Internet. As I mentioned, we're not only aggregating jobs from the major job board, we're also monitoring a lot of individual company core site. When they come here, instead of giving them a search bar, we actually do the match based solution. They just upload their resumes and answer a couple of questions with AI. And you can see we instantly present them the jobs. We think they have the highest chance to get an interview. On the right side, we show them a score that's indicator when the HR look at your resume, how strong they think your background is aligned with the jd. And then when they hover on that, we further give them details on your experience level match, skill match and industry levels. And also you can see all those jobs are very fresh like within hours. Because definitely when you apply, when the job was posted, the earlier you apply, I think you get edge over the others. And on the right we have Orion is basically your AI job search copilot. You can chat this for any jobs. For example, I can click ask Orion for any jobs. I can ask any questions job, right. Basically research about this company. You can ask what's the funding history of this company? What is the business of this company. More importantly, I can ask okay, for this job instead of a score, tell me why I'm a good fit for this job. So the Orion will give you like very explicit explanations on like why they think it's good. You are a good fit in terms of experience, seniorities, education, core skills, right? In this case, you know a lot of job seekers they don't get many interview because they just randomly apply jobs which is open. But for job right we want to make sure the job you applied you are qualified for that. So that's number one. So number two as you mentioned, resumes, right. Not many users know how to write a professional resumes and even ATS friendly. So for the resume after you upload your resume we're first do analysis across your upon your current resume versions. Let's say for this resume, right as all the resume writers do, we basically first analyze your resume, give you a report. For example for this resume we score it as a b. We find a 6 urgent fix and this is the summary the categories of issues we actually find. And if you are curious, okay, so can you help me to optimize it. That's definitely what we can do, right? So you can go to the working experience. We tell you, okay, each of the bullet points for example for this one and obviously we think this bullet points lack of achievements and why this is important issues and how we think the AI think it should fix it. And then it directly give you a suggested versions and you can use if you are not satisfied with it or you can add a bit more information. You can rewrite it with AI just like how you engage with AI and after that you submit the version. It's just so easy for you to just go through all the issues and instantly when you realize it, reanalyze it. You get like score a resume for your job applications.

Got it. So it really helps you make your resume. If you don't have a resume, let's say you just have a LinkedIn page with your, with your jobs listed there. Does it make a resume out of that? Does it work with the, with the LinkedIn?

Yeah, definitely. So actually in the onboarding process, if you don't have a resume, you can just let us know your LinkedIn, your, the. The. Your LinkedIn profile address and then our AI will just to your current LinkedIn profile put it in here. And then it will also do the analysis and give you suggestions like help you to write a summaries. If you don't have a summary, it will help you to write it for your working experience. It will make a suggestion basically working with you and to quickly build up a good resume.

No, that's very helpful. Anything else we should see here? That. Yeah, so the power of AI particularly.

Yeah, as I mentioned, as you mentioned. Right, so. So we not only basically optimize your base version of a resume for every single job, we actually customize it for you. So and as I mentioned, right if you just. Let's assume you are just interested about this job and we actually just can. You can just come here and you can just generate a customer resume first. Our AI will give you a side by side comparison between your current resumes and this job. And in terms of top job titles, years and experience, industry skills, what are the match which didn't, which are not. And then when you just click okay, I want to optimize it. The AI will first confirm with you with the missing skills in your resume because it can make up data, right. So we want to make sure so all the information is truthful when you just click on the scales you actually have. And then so you just click generate new resumes after 10 seconds show a tailored resume is out compared with these kind of most of the resume tools we actually don't only just tailor your keywords, we actually just try to render your resume as professional as possible. We try to optimize your summaries, optimize your latest bullet points. There's a six second rule. Right. So for a recruiter for a given the resume they Only have a 6 second to read your resume. We want to make sure within that six seconds all the content, especially the first page of the resume can definitely draw their attention and then they can download it and then directly send it out. The last one feature is kind of unique. We build it, we call it insider connection. You know people everyone use LinkedIn for network. But the thing is you might not have a lot of like your friends on LinkedIn. Given the job you might want to find the peers or like hiring managers or potential the colleagues or someone who maybe have a like second tier or third tier relationship with you. So you can send a message for referrals and job right. Actually automate all the networking effort. Given this job we actually try to find you the hiring managers or the peers for that target positions. Even more than that because we know your school, your work experience based on resume. Actually for this job we actually find your ex colleagues working at Adobe and then so that it just makes the connection more personalized. And then beside that we help you to actually to find the work email of that contact and then even draft the personalized email for you. Right. So you can see we directly draft a personalized email based on work experience and to that position. Yeah. So all you need to is just to send the email or connect people on LinkedIn. So after using this feature a lot of our user actually get a lot of feedback directly from their LinkedIn connections. Yeah. So those are the things I think there's a lot more features you can hear. You can basically ask Orion and get information, 360 degree information about the company, about their own history. Leadership team. The latest news, you don't need to Google search anymore. Bring all the information.

Everybody's trying to kill Google.

Yeah, we're not killing it. We're relying on the information providing by them. But yeah, we definitely make the job search so much more efficient. Yeah.

Now this is really really helpful for people. So thank you very much. Give me how long did this take you to do because this looks really really well done. It's not just a wrapper on chat GPT, you know, quick and dirty kind of company.

Yeah, you are right. I Think it's. It's pretty. Actually the, the tech stack behind the products is pretty challenging. We're actually. Instead of just using ChatGPT, I mean building features as a wrapper of ChatGPT, we. We actually try to use larger than models in the matching because it's pretty. When you look at the job recommend to you, we want you to feel like, okay, this matching is similar like the quality a human expert deliver to you. So Job right was founded just like last June, so a little bit more than like one year and four. Four months. But we have a pretty great team with several AI engineers. They're very good at recommendation system machine learnings and larger models. And yeah, thanks for their expertise. And then. Yeah, so we can build this product very quickly.

Yeah, yeah, that's really cool. Tell me, how did you get funded? How did you get the funding to do this?

Yeah, so last week. Sorry, last year. So when we started. Because the jobrite is my second startup. So actually so. So when we found this company, there's already several investors, they have been following with me and keep connected in the past several years. They think this is really cool idea because they also hire people so they know how efficient it is for the hiring side as well. So they think this is definitely a great opportunity. So they invested give us a seed round of fundings last year.

Yeah, this is really cool. Tell me a little bit about the technology underneath. What kind of models are you using and how are you looking at AI? Because you're actually using it at a very advanced level, I think in terms of building a product, building a company. Yeah.

First, I think the major reason we decided to build Job Right is I think before the breakthrough of large element models, I think all the past technology was it's almost impossible for the past technology to actually do the matching like a human. It just only do keywords matching. Right. But after ChatGPT was out, I was a user. I was super shocked by its experience, especially its reasoning ability. Like a human. Think like a human. That was the moment. I think, okay, there's a new opportunity, new solution which was not possible before and now it's possible right now.

So.

Yeah, so I think heavily Job Red Stack Stack relies on the larger models we actually start to use.

Do you have your favorite.

Sorry?

Do you have a favorite that you're using? Are you using Llama or Mistral or.

Yeah, we actually. Yeah, so it's a hybrid and especially like we start to use the public models like GPT4 and especially. Yeah. But for some of the and use the GPT4 some for some kind of standard data processing. Because we have a lot of jobs and a lot of information we aggregate from the web. We need you the large end model to basically parse with the information into the text or the feature, the embedding which can be used by the recommendation systems. But the more we build, we've actually found there's a limitations with the public models no matter it's a GPT or cloud because it's very all around large models but it doesn't have the expertise into this area for example for the resumes. And if you now give a resume to the ChatGPT and ask it to give you suggestions, it can give you some, but it's very general, generic and wordy. And then so basically we decide, okay, so we might want to fine tune a model which can be used for a specific case then. Yeah, so we use open source models like Llama, which is really awesome as well. And then we fine tune the model using our in house data as well.

Very cool. So you're probably pretty expert at setting up databases too because it sounds like you're grabbing a lot of data from all over the place and building that into a little model. I'm sorry, are you using a vector database to personalize the thing for me? Because it seems very personalized to me because it knows my resume and my socials and all that.

Yeah, definitely. So basically what we do is on the both side, once we have the information, all the raw data and then we use a larger than model to process it and build the embedding which is like thousands of dimensional vectors and then store it in the vector database. And then on the job seeker side, the same when they upload their resume, when they engage with AI, we are consistently getting more understanding about them. And then we optimize and build the profile in the. The the vectors for. To represent their profile and then we do the matching using. Yeah. So under.

Yeah, yeah. How are you finding work building a product and a company with AI so deeply embedded, it's not just bolted on, you know, what do you hope the AI industry does better?

Yeah, definitely. And first, thanks for the breakthrough of larger model. It gave us the new opportunities to build something really cool. Especially after the mobile era. Right. So this is definitely another round of exciting technology. So innovations I think. Definitely. So we're on the application side as I mentioned. So when I decided to work on job, right it's not like okay, the AI comes, I need to find a great opportunity, we can leverage it actually originally I already feel dependent as a job seeker and hiring even before ChatGPT was out. So I just didn't get the perfect solution to solve this until it was out. So I think as a founder definitely you need to be problem driven instead of just like chasing for the. The hype of the technology and just. And in this case you might, you might not create a lot of values and we do see we do have a. Because I mean as an application small startup there's no way we can build the foundational systems. I think in terms of our expectation for the large end models, definitely there's top two things. One thing is I think the goal is to build an engine which can do the reasoning as. As well as the human expert. But Maybe the current GPT4 or the latest models and I think they're in term of for the reasonabilities there's definitely a big room for improvements. Right. So maybe it's not like a high school students but we're aiming, we're looking at maybe GPT5 can be a college student or something like that. Right. So the reasonability is definitely limiting the effectiveness of our mesh engine. The second is how the largest model can give more truthful information. Sometimes we can make illusions problems. Right. So especially you're talking about resume. You definitely want to make sure all the information on your resume is truthful. You are not making mistakes that will bring them troubles. So I think that I would say that will be the two things we're definitely considering monitoring the latest technology in the. In the tech. For example, I think the latest 01 is awesome because it can. It's a bit slower than the GPT 4.0 but it can do much better jobs on syncing. That is actually the things we're waiting for.

That's cool. One last question. What are you learning about the job process that maybe shocked you by building this company? Because you're having to do a lot of working with customers and with the database and prompting. You probably had to write some custom prompts for this or fine tune your engine. I bet you learned a lot about getting a job.

Yeah, I think the number one thing I would say in the past one year I learned is you know, as a. As I mean for any professionals, I mean we might only know 30 to 50 companies at most. Right. The biggest one, the hottest AI startups but when we actually because we have a lot of data about the jobs they there job seeker feel like it's super tough right now in the market. True. Truly compared with four years ago or like, it's definitely tougher. But when you look at the data from the recruiting side, there are actually a lot more opportunities than what job seekers think. But the thing is, because they don't, if they go to LinkedIn, they, when the LinkedIn give them jobs, a hundred jobs, they might only know 10 of them. That means for the rest, 90 of the company, maybe they're great, it's suitable for them, they can get an interview. But the things they don't understand about the industry, they don't think, okay, as a product manager, actually I'm suitable for another job title which can also leverage my skills. That's the number one lesson we learn. I think most professionals, they're very talented, but when they do the job search, they're very limited to their understanding about this market. There's actually a lot like rising tech, rising industries and positionings, they can be a perfect fit. So that's basically why when we do the recommendation, we don't do the match based on titles, we do the match based on skills, based on their strengths. So we can discover the jobs they never thought might be a super hotel. That may be the one thing I want to share.

That's awesome. Well, thank you so much for helping job seekers because there's a lot of us out there looking for jobs. Right. Where do I find out more about you and about the company? Where do we try this?

Yeah, so jobwrite is a free tool. Everyone can just give it a try to Jobrite AI and for me, we stay on LinkedIn. Anyone who want to have a chat with us and give us feedback and just shoot me a connection on LinkedIn, it will be the best way.

Thank you so much.

Yeah, thanks for the time. Yeah. So reminding me to share about job right here as well.