Modified End To End Resume ATS Tracking LLM Project With Google Gemini Pro

hello all my name is kushak and welcome to my YouTube channel so guys just a couple of videos back in my channel I had actually created an end resume application tracking system and over there you know the kind of implementation I had specifically done uh had lot of setup issues specifically with respect to one of the executables like popular okay because we were using one Library which I will again show you in this video I'm going to modify that particular project in a better way so that you don't get any kind of issues right and the way that I'm specifically going to use will be very super easy because in that project what I did is that I took the PDF converted into an image and from that image I'm actually getting bytes okay now from converting a PDF to image had some of the setup issues with respect to this popular where they'll say that PDF file not found something like that but in this video what I will do is that I'll take that PDF I'll extract the text and then probably I'll go ahead and create my ATS system so uh just to go ahead this was the video that I'm actually talking about end to endend resume application tracking system and if I probably go ahead and see in all the comments okay so you here you'll be able to see that uh uh there are some errors like this actus error request failed with status 403 some of the other errors are there over here see over here hi Krish I'm facing issue with pdf2 image so this was the library that we specifically used now let me go ahead and uh let me try to write a modified version of this particular project with respect to the code so again I will probably start completely from scratch so please make sure that you try this and this this way that we have specifically done is pretty much easy when compared to the previous way that we had actually done again this was suggested by one of my student you know which is super amazing to see and that really makes me proud right Chris you can probably go ahead and do this specific way right so considering that I'm going to go ahead and probably do in that way itself so that you don't get any issues so first of all I will go to my terminal okay again we are going to create this end to endend modified version itself so I will quickly go ahead and create my environment now already in my previous video also I've shown you how you can create different different python environments in a data science project or in any kind of project itself so I will probably go ahead with cond I've already told you what is the use of cond itself uh I will use Python dou equal to 3.10 and then I'll also give y so once this particular installation probably takes place and here you'll be able to see that all the requirement. TX I mean all the environment that is basically required will be created over here okay so quickly let's see why it is taking time okay that is the reason right I did not write the command properly it's okay so I will give like this great so here you can probably see now everything the installation has been taken place and it is taking place and here you'll be also able to see the V andv environment along with this I'll will be requiring environment file along with this I will be requiring my requirement do txt file okay uh till this installation is taking place what all specific libraries we want from the requirement. txt I'm just going to copy it and paste it and this time I'm not going to use pdf2 image okay I will be using pi pdf2 because with this you can actually read all the text that is specifically there in the PDF and that is what we require to summarize that entire document right so I will be using this library and along with this what I will do quickly I will first of all activate so I will go ahead and write cond activate VNV okay this all steps I'm going a little bit fast because I've shown you already if you have not seen those videos please go ahead and see it because step by step I've done it I don't want to waste much time in this okay so now I will go ahead and install after activating I'm going to install the requirement. txt so so once this is getting installed till then I will also go ahead and update myb I've already created my Google API key right how do you create a Google API key go to the maker suit I'll just write maker suit website okay which will actually help you to create it so this is my Google API key which is currently working I'll remove this because I don't want it okay now once this installation of the requirement. txt will take place I will start my coding now what I will do I will quickly go ahead and create my app. py file again as I said what is my main aim and what is the mistake that I specifically done in the last session okay and why you were facing many issues suppose let's say this is my PDF okay so let's say this is my PDF okay initially we were converting this to image right and when we are converting this to image to get the bytes here we are using a library which is called as PDF to image okay so this was the library that was used and for this the dependency was that we need to install and set up popular right and for Windows for Mac for Linux there is a different way of setting this up okay after that we converting this into bytes that basically means we are taking the info and then we are probably going ahead with the API which API Google jini pro and finally we are getting the response and obviously this usually takes time now what we are doing the in the modified version and this is the approach that we are following I will take the PDF convert this into text right and already we have that feature from PI PDF 2 right and then I will take this text other than this I don't need to even convert that into some other bytes itself I will take this text and probably hit my API with respect to some prom template and here I will go ahead and get my response so this is the approach this is the approach that I am going to follow right approach to so that is what I'm actually planning to do over here now let me go ahead and probably let's see the installation is done everything looks good now I will go to my app.py file and start writing the code and today what I'm actually going to show in this particular video is that how you can create a prom prompt template by providing parameters inside that prompt okay so that also I'll be showing I've never shown that with the help of Google generative AI so I will go ahead and write import streamlet as a steam okay okay now this is done import streamlet okay as SD along with this I will go ahead and write import Google dot Google dot generative AI as gen AI okay let me just go ahead and write I don't know why this is not giving me any suggestion as such yeah now it's coming but perfect let's see whether there is some issues over here okay perfect just a second okay I will just go ahead and write the code now open my terminal new terminal command prompt everything looks good V andv environment if I probably go ahead and see it okay import streamlet as St then import generative AI as sorry google. generative AI Google do generative AIS gen a done okay along with this I'm also going to import OS because I I I'll show you where I'll be using import OS then along with this I will go ahead and write Pi PD pf2 okay as PDF now this is the library that I'm going to specifically use and that is how I'm solving my problems okay as PDF oh come on I have to clean all my extensions I've used lot of extensions so that is the reason it is probably coming from with respect to this you know from do EnV I'm going to import load uncore Dov okay so this is also done I will go ahead and initialize load. EnV so that I call all my load all the environment variables okay done so this is done okay now once we are doing all these things uh the next thing that we really need to do is that configure my configure my API key API uncore key is equal to os. get EnV and here I'm going to specifically use my Google API key for my environment variable okay done once this is done now we will write gini Pro response okay so let me go ahead and write my definition so here I will say getor Jore response and this will basically be my input after taking my input first of all I will load my model so I will go ahead and write um sorry gen AI do generative generative model and here I'm going to specifically give my model that is called as jini Pro okay before I used to use gin Pro Vision since I was converting a PDF to image now while converting PDF to image I used to use use a another Library okay so gin Pro is done over here then I will probably get the response and I will go ahead and write model dot generate underscore content and here let me go ahead and write this input okay whatever input I'm getting and along with this what I will do I will go ahead and write response return response. text done come on I will fix all my extension I don't know there a lot of extensions I think I installed because of that so many problems are basically coming let me just save it okay this looks good uh response this is there now the second function that I'm specifically going to write to convert or to from that PDF right I really need to extract the image so what I will do I will just go ahead and write my another function input uncore PDF unor text and here I'm just going to give my uploaded file okay now first thing first as soon as this uploaded file comes over here I'll be using I'll be creating a variable which is called as reader PDF do PDF reader okay PDF reader and here I'm actually going to give my uploaded file and this is how you specifically read it okay uh once you get this reader object I'm going to create a text variable and then explore in every page in this reader. pages right so there will be multiple pages in this particular PDF right so let's say reader and here I'll say length of because I need to iterate through every pages so let me go ahead and write reader. Pages uh and then I will write colon uh the next thing that I will be doing is that I will extract all the information from this particular pages so let me go ahead and write page is equal to reader. Pages reader. pages and I'll give the page number over here okay so this will be the Page information and then I will go ahead and put the entire content from that particular page to this text variable and then I'll also convert that into string so in order to extract that text I have to use this function called as extractor text okay so this is how simple it looks uh and then finally I will go ahead and return my text so over here you can see this is responsible for taking the PDF and extracting the text and then we can give take that text along with my prompt template and I can get the Google gin Pro response so two important steps over here right now is the most important thing is that about my prom template now prom template is something that you can design as you like you know it's more about verifying things so here you can probably see I've created my prompt template and this prompt template looks amazing improved prompt I have say hey act like a skilled or very expens experienced ATS application tracking syst with a deep understanding of tech field software engineering data science data analyst and Big Data engineer your task is to evaluate the resume based on the job description you must consider the job market is very competitive and you should provide the best assistance for improving the resumés assign the percentage matching based on the JD and the missing keywords with high accuracy so here is what I've given I've given the resume as text okay which from this particular function I'm getting the text and then I'm also getting the JD right description which I will probably put it in a field I want the response in one single string having the structure JD match missing keywords like what are the missing keywords based on the job description and profile summary so this is my entire input prompt template now let me go ahead and quickly create the streamlet app for you and this step I hope everybody is very much similar with hd. title is smart as uh ATS std. text improve your resume ATS I've taken this text area to probably put the JD and that JD is probably going over here right and this text will be probably coming from this particular function right and then I've also Ed an upload file I'm saying upload a resume to type is PDF please upload the resume and this is submit button now if I probably do the submission button right then what will happen I will go ahead and write if uploaded if uploaded file is not none okay if uploaded file is not none then the next step what I will go ahead and write text is equal to input input PDF text because I want to convert my uploaded file into text right so here I will basically give my uploaded file okay after getting the text I will go ahead and call my get Gemini response and here I will give my entire text okay uh sorry not text because I need to give this input prompt because text will get allocated over here right so here I will probably give my input prompt so this entire input prompt will be going to the Google gy and it will tell okay these are the thing that you really need to act and this is how your resume looks like JD looks like okay so once you specifically do this uh here I'm going to get the response okay and along with this what I will go ahead and write St do subheader SUB header and here I'm going to display the response okay done now I think everything should work fine uh very simple only the thing that I've done is that I've used this Pi PDF 2 and implemented everything that I really wanted okay the previous one had lot of issues with respect to the setup so obviously I got a lot of message from many people itself but anyhow I'll be fixing this okay now let me go ahead and write streamlet run app.py now I think it should be working let me just open LinkedIn right really fascinated with LinkedIn guys okay really really fascinated now with LinkedIn you will be able to see different different job description let's say over here there is a lot of job updates in LinkedIn and if you have a good ATM dynamically resume trust me it will help you guys many people they apply for jobs the main thing what they do is that they have the old resume they don't even have a look onto the JD right right and they start applying for it and then they'll say that hey Krish I'm not getting any job calls right what is the problem over there just try to think okay you need to always make sure that your resume is ATS proof right that basically means the ATS system should be able to short list your resume if it is not matching with the JD it will not be going to happen lot of MNC companies are there a lot of other companies are there who will actually use this ATS system now I will go ahead and browse my resume so so guys now I have have uploaded the job description and the resume now let's see whether this is going to work or not now once I probably submit it here you'll be able to see that it's working right now here it shows that JD match is 75% missing keyword C C++ Java no SQL machine learning data visualization some information that you can specifically write with respect to this see there will be multiple things that may be missing from your resume if you have not learned that specific thing it's okay skip it okay or at least try to write some information that you know about those keywords right because at the end of the day if your resume gets shortlisted first then whatever interviews are there right let's say over there if you say that I don't know this I don't know that right then they may ignore those things right but if you are maximumly telling about job uh about your projects that you have specifically done the implementation that you have done and if they like it trust me you have a very high probability to get in that particular job selected right so I hope you like this particular video this was it from my side guys now I hope you'll not face any kind of Errors uh this is it for my side I'll see you in the next video have a great day thank you one all take care bye-bye

Channel: Krish Naik

Views: 49,029

Keywords: yt:cc=on, generativeai, large language model, google gemini pro, google gemini pro vision, ATS system, resume tracking system, ATS app using google gemini, pypdfreader, krish naik generativeai

Id: VZOnp2YpY8Q

Length: 18min 8sec (1088 seconds)

Published: Sun Jan 21 2024