Transcript for transcript_153InEzjKKhuiyoIHiDxYtOpPO53fwkcu_1_1749_181fab149276_1.txt

transcript_153InEzjKKhuiyoIHiDxYtOpPO53fwkcu_1_1749_181fab149276_1.txt

[00:00:00]  We want to create a supply chain of urea demand.
[00:00:04]  So that was our project and we reached in final in college days.
[00:00:12]  Okay.
[00:00:14]  And tell me about your technical skills.
[00:00:17]  So in technical skills, I did Java, DSA and core computer science project and later on it was separated to be a modern stack and
[00:00:27]  in recent days I did work on like a data engineering, one of data engineering project like a Mandelian architectures like a
[00:00:35]  Bronze layer, silver layer and gold layer using Spark and Kafka.
[00:00:39]  So this is my technical.
[00:00:42]  Sorry, what you have mentioned?
[00:00:44]  Spark and?
[00:00:46]  Kafka.
[00:00:47]  Kafka.
[00:00:48]  Okay.
[00:00:48]  Could you show me any websites that you have, any projects that you have worked on this thing?
[00:00:56]  So, I still like I have done like
[00:01:04]  more the theoretical or more the working on different kind of like a lecture level, but I did not create such any like a GitHub kind of project that is a fully completed project.
[00:01:19]  Okay, that's fine.
[00:01:20]  I just want to know your technicalities.
[00:01:21]  That's it.
[00:01:23]  Yeah.
[00:01:24]  Okay.
[00:01:29]  So you have a difference in MySQL, MongoDB and Postgres, right?
[00:01:37]  Yeah.
[00:01:37]  So what is the difference between Mongo and sorry, MySQL and Postgres?
[00:01:43]  So MySQL and Postgres both are databases, both can we use a primary databases, but Postgres is a open source, Postgres is a open source and
[00:01:57]  both are open source databases, but the differently main design philosophy like MySQL is a simple, faster and ready heavy workload.
[00:02:06]  Postgres are most advantage feature like better suited for complex queries and data integrity.
[00:02:13]  And if we talk about architecture level, so MySQL focus on speed and simplicity.
[00:02:20]  And in post-Grace SQL focus on standard compliance and extensibility.
[00:02:26]  And they both follow a seat property like atomicity.
[00:02:31]  Consistency and durability and isolation and performance is my sequel is faster but post grace also faster but post grace used to
[00:02:45]  handle heavy data like a bulk amount of data okay okay
[00:02:52]  You know deployment as well?
[00:02:54]  Yeah.
[00:02:55]  You know deployment also?
[00:02:58]  Yeah, I deployed one of my project that was a currency converter.
[00:03:02]  I was using an API of currency converter that I deploy my project on Versal.
[00:03:08]  But in your resume, you have mentioned EC2S3, right?
[00:03:11]  Yeah, I have mentioned, I did not mention Versal.
[00:03:14]  I have worked on EC2 instance AWS.
[00:03:18]  You know how to deploy it?
[00:03:19]  Yeah.
[00:03:20]  You know how to deploy it?
[00:03:23]  Like, yeah, I know how to deploy, but there was a problem when I was deploying at that time.
[00:03:28]  And again, it was not working properly.
[00:03:31]  So I was no like a full kind of experience that I can do very well.
[00:03:38]  Okay, you know Docker?
[00:03:40]  Yeah, Docker.
[00:03:41]  Yeah, Docker is used for containerization.
[00:03:44]  So I used for you.
[00:03:46]  We use Docker.
[00:03:48]  Yeah.
[00:03:49]  We use Docker.
[00:03:51]  Yeah.
[00:03:52]  Okay.
[00:03:54]  So what is root folder in that?
[00:03:58]  root folder in docker so i would say like
[00:04:08]  like a root folder like refers to a file system like inside a container like not in
[00:04:16]  local machine root directory like every docker container is on isolated file system so file system start from like a slash root so just like it is
[00:04:29]  It's created a Docker image layer.
[00:04:32]  So I think I know about it.
[00:04:36]  Okay.
[00:04:38]  Okay.
[00:04:39]  Am I right or wrong?
[00:04:42]  Most probably.
[00:04:45]  What is GCP?
[00:04:48]  Yeah.
[00:04:49]  What is GCP?
[00:04:52]  GCP, Google Cloud Platform.
[00:04:53]  Yeah, it's a computing platform by Google that provides on-demand infrastructures and services like compute, storage, database, networking.
[00:05:06]  Okay, you have done anything on machine learning?
[00:05:10]  I created a MLOps project on machine learning that ML flow and cube flow and yeah, on a theory basis, I have done many machine learning like a supervised and supervised reinforcement something.
[00:05:23]  What is reinforcement learning?
[00:05:26]  In reinforcement learning, we train data gradually, like a step by step.
[00:05:31]  We train data like it's a type of machine learning.
[00:05:36]  Like we give a training data and we test the data.
[00:05:41]  Again, we give again training data.
[00:05:43]  Step by step, we are giving the data.
[00:05:48]  Following like a error should be a less.
[00:05:51]  So our main concern is to reduce the error of that is differ from accuracy.
[00:05:58]  So we are giving step by step.
[00:06:01]  Okay, what is CNN?
[00:06:04]  Yeah, CNN stand for computational neural network.
[00:06:08]  So it is a deep learning concept and it is a type of deep learning model designed processing grid like data structure by as image automatically learning a special feature using convolutional operations.
[00:06:22]  So traditional neural network like CNN is like a traditional neural network we use.
[00:06:30]  So capture like in deep learning, we use for capturing local patterns and preserving separate structures.
[00:06:40]  Okay.
[00:06:42]  Okay.
[00:06:43]  Could you show me any previous project that you have worked in ML or deep learning or something?
[00:06:51]  One of project that I worked but it is not in like how do I was send it on my LinkedIn
[00:07:07]  In GitHub, I did not.
[00:07:09]  I think my account is not working for.
[00:07:13]  Let me see if I can.
[00:08:10]  Yeah.
[00:08:12]  So one of these project I was doing in Malops, like we were creating a random forest and logic recreation.
[00:08:28]  I have shared my screen.
[00:08:32]  It is visible.
[00:08:33]  Yeah, it's visible.
[00:08:34]  It's visible.
[00:08:35]  Yeah.
[00:08:36]  So I have worked on it and yeah, random forest and logistic and we are
[00:08:44]  doing for accuracy for it.
[00:08:48]  So yeah, so I was like creating Melov's pipeline, but I did not deploy it on GitHub.
[00:08:55]  So I can't tell you like this.
[00:08:58]  I have code in my local machine.
[00:09:01]  That's fine.
[00:09:02]  What is the purpose of this project?
[00:09:05]  Yeah, like a purpose of this project, creating a machine learning operation, like we have different kind of model and we want to predict for accuracy,
[00:09:15]  like a logistic regression and random forest both comes in supervised machine learning.
[00:09:21]  So we create a model in supervised machine learning.
[00:09:24]  We have two type of first, we have a training data and testing data and we have both input and output both.
[00:09:31]  So we give the model input and output both and we predict the accuracy and we create the model.
[00:09:39]  So the main agenda was to create that particular model which predict more accurate result.
[00:09:49]  Using this kind of algorithm.
[00:09:54]  Okay.
[00:09:55]  Yeah.
[00:09:56]  Okay.
[00:09:57]  So, okay, that's it.
[00:10:00]  I understand.
[00:10:05]  Any other project?
[00:10:07]  Yeah, one of project that I was working like if you saw it is a currency converted as I mentioned previous.
[00:10:17]  So it was like a project I was working.
[00:10:21]  Yeah, it is a it is a deployed on Versal.
[00:10:25]  And that's fine.
[00:10:27]  This project, I understand.
[00:10:29]  Other than this.
[00:10:32]  Yeah, other than this, I have one more project in GitHub.
[00:10:37]  So that project is like what I mentioned in my internship, like a wonderful project based on man.
[00:10:46]  Show me.
[00:10:51]  Just be doing.
[00:10:53]  Okay.
[00:11:15]  No, I think I have removed that project, but one of one more project that I have to show, like, if I did not remove it from the my GitHub account.
[00:11:25]  So sorry for this also, I think I have removed from it.
[00:11:30]  Why did you remove it?
[00:11:32]  So I was thinking like it is not full-fledged.
[00:11:36]  It was not working properly.
[00:11:37]  I have to move it from scratch again.
[00:11:40]  So I removed it.
[00:11:41]  So I think that is it.
[00:11:43]  Will you show me in your local?
[00:11:47]  Show me my local.
[00:11:49]  I want to see the code.
[00:11:51]  You want to see the code.
[00:11:53]  So I think I am using second laptop.
[00:11:56]  I have another laptop in which I created on project.
[00:12:00]  So I think in this laptop, there is no sorry for inconvenience.
[00:12:16]  Sorry, I don't have a project on this laptop.
[00:12:21]  Okay, fine, fine.
[00:12:23]  Okay.
[00:12:29]  Okay, I'll give you a use case.
[00:12:30]  Okay, just give me the solution for that particular problem statement.
[00:12:37]  Okay.
[00:12:37]  Okay.
[00:12:39]  So think that open chat GPT.
[00:12:44]  Open chat GPT.
[00:12:47]  Yeah.
[00:12:51]  Okay.
[00:12:52]  So if you type anything, it will type who am I?
[00:13:03]  Just log in, log in.
[00:13:07]  Sorry.
[00:13:19]  Yeah.
[00:13:20]  Now type, who am I?
[00:13:23]  Who am I?
[00:13:24]  No, C-H-O-O.
[00:13:25]  Sorry, W-H-O-O.
[00:13:30]  Space, space, space.
[00:13:33]  Who am I?
[00:13:35]  Yeah.
[00:13:36]  Thank you.
[00:13:37]  And tomorrow?
[00:13:44]  Yeah.
[00:13:46]  See, you're asking about yourself and it is giving you the particular data, right?
[00:13:51]  About yourself.
[00:13:53]  How it is giving you, you know that.
[00:13:57]  Like I am not understanding what you want to say.
[00:14:00]  Sorry.
[00:14:00]  You have just typed who am I?
[00:14:02]  Yeah.
[00:14:03]  Okay.
[00:14:03]  It is giving you a result as you are Shivam Divedi and you are learning these kind of things.
[00:14:09]  Right.
[00:14:10]  How did ChatGPT know that you are Shivam Divedi?
[00:14:15]  Because I have given my previous for the same login account, I have given my details, I have chat and they are using my previous data while when a user asks chat GPT about like.
[00:14:30]  When a user about they use their previous data and how they are logging
[00:14:35]  they are working and i have like a user as chat gpd their own details so system does not remember everything by default but instead it use combination of like a const level memory system and
[00:14:50]  llm reasoning so used to generate that kind of answer because i have used previous time and
[00:14:57]  So using this LLM reasoning to generate answer.
[00:15:01]  So because of LLM reasoning, they are giving me answer that kind of.
[00:15:07]  So you are telling me that your data is stored in chat GPT, right?
[00:15:15]  I am not telling that my data is stored in chair GPT, but my like, how do I say?
[00:15:27]  No, I'm not saying that data is stored in case like, let me think just a second.
[00:15:43]  So personal data is not inherently stored by chat GPT as I know the rule in the way databases use
[00:15:53]  user profile, I think, and unless system explicitly implement memory or presentation.
[00:16:00]  So.
[00:16:01]  I think they use for user profile.
[00:16:04]  So my user profile is same and they know my they have stored my like a hierarchical behavior.
[00:16:15]  So I think that is going and that's why this is answer is generating.
[00:16:22]  Okay, that's I understand.
[00:16:24]  But is ChatGPT storing the data that you're typing each and every day?
[00:16:34]  I don't know.
[00:16:36]  Yeah, I think, yeah.
[00:16:42]  Like, yeah, it's definitely stored my data and use for that data for the all, but not my personal data.
[00:16:53]  I'm not asking about your personal data.
[00:16:55]  See, the chat GPT is giving you a result as Shivam Gawadi.
[00:16:59]  It's your name.
[00:17:00]  Okay, fine.
[00:17:01]  Yeah.
[00:17:01]  Coming to it, let's assume that it took your name from Gmail.
[00:17:06]  Okay.
[00:17:07]  Coming to the next point, you are currently learning Python, NumPy next.
[00:17:12]  Yeah.
[00:17:12]  Okay.
[00:17:13]  Working on generative A major project.
[00:17:16]  Yeah.
[00:17:16]  After that, preparing for software tech opportunities, possibly Bangalore.
[00:17:21]  Yeah.
[00:17:21]  How did chat GPT know that you are doing these things?
[00:17:25]  Because in daily life, I daily use to give that kind of data.
[00:17:30]  So it is storing your data or is this randomly popping up data from your previous chat?
[00:17:37]  By seeing this, I can say it is storing my data and it is processing my data like a pre-trained knowledge training data using as a training data on my input user training data and input prompt like.
[00:17:52]  Okay, so you think that chat GPT is storing your data.
[00:17:55]  Let's assume that.
[00:17:57]  So is it good thing that chat GPT is storing your data?
[00:18:02]  So if I ask a question about like a currently learning Python, generative AI and tech opportunities, if they are storing that kind of data on a journal perspective that I don't have any kind of problem.
[00:18:15]  Instead of they are storing my personal data, if I ask something about my personal things.
[00:18:20]  So if I have personal like a health issue problem.
[00:18:23]  So if they are like a storing like as an individual, like I have a personal problem.
[00:18:29]  So if their problem, generalize the problem and they store that particular problem.
[00:18:34]  So I don't have any problem.
[00:18:36]  So they might be help for others by processing this data.
[00:18:41]  Okay.
[00:18:42]  Okay.
[00:18:43]  So my question is, can we build a similar kind of application?
[00:18:48]  Similar kind of system?
[00:18:52]  So as far as similar kind of system.
[00:19:02]  I would say...
[00:19:07]  Yeah, definitely we can build, but...
[00:19:13]  Like a personalized like a stat gpt like a system like personalized on users on data but
[00:19:23]  like we need to reach architecture like llm memory and retrieval like the practical approach is like
[00:19:34]  General purpose LLM, we need to LLM, we have to working on LLM.
[00:19:39]  So after it's like a high level architecture, definitely we can create.
[00:19:45]  Okay, okay.
[00:19:46]  So you know what is GCP Vortex?
[00:19:50]  GCP vertex.
[00:19:52]  Yeah.
[00:19:54]  So GCP vertex like GCP.
[00:20:00]  Pivortex.
[00:20:02]  Have you heard about that thing?
[00:20:03]  Yeah, I heard about it, but I'm thinking about just a definition.
[00:20:07]  How do I say to you?
[00:20:09]  So, like I heard about Vertex AI, that is a fully managed machine learning platform on GCP.
[00:20:20]  That allow to build, train, deploy and manage ML models.
[00:20:25]  So I heard about it like a ML pipeline require multiple tools for training, deployment, monitoring.
[00:20:33]  So GCP Vertex working on that for managing kind of things.
[00:20:42]  Okay, okay, sure.
[00:20:45]  That's fine.
[00:20:47]  What is embeddings?
[00:20:49]  Embedding.
[00:20:51]  Yeah.
[00:20:52]  So embedding is like a numerical vector representation of data.
[00:21:02]  Like such as similar items are placed closer in high dimensionality space.
[00:21:09]  And yeah, principal one of principal component analysis that I also work, I have also studied about in deep learning.
[00:21:18]  So the dimensionality reduction is also one of and computer cannot understand raw text meaningfully.
[00:21:25]  So embedding converts text into numbers like.
[00:21:30]  we have dog and like example we have a dog and another kind of dog like this other breed so like a similar vector but
[00:21:40]  dog if we compare to dog with other just animals like not similar breed or dog with be a cat or be a lion so it's a very different vector so embedding work on that
[00:21:55]  Okay.
[00:21:56]  What is metadata?
[00:22:00]  Meta data.
[00:22:04]  like metadata is data that describes other data is provide context structures meaning and like
[00:22:15]  actual data like like i have a data like i have pdf so i this data is actual data but metadata is like what is pdf file name or what i what is size what is
[00:22:30]  which date pdf is created so this are kind of like i had there is many kind of metadata like a descriptive metadata like help identify the data
[00:22:46]  Okay.
[00:22:47]  Okay.
[00:22:48]  You're currently in which state?
[00:22:51]  So I'm currently in Hyderabad, Gondapur.
[00:22:55]  Okay, what are you doing in Hyderabad?
[00:22:58]  Yeah, I am doing self-study and I am doing creating my project since I have completed my graduation.
[00:23:04]  So I'm looking for opportunity.
[00:23:06]  My friends are working in MNC.
[00:23:09]  So I am living with them.
[00:23:17]  Okay.
[00:23:19]  Sorry for the grammar.
[00:23:21]  I have such a lack of like, you know, that's fine.
[00:23:26]  I understand.
[00:23:28]  Okay.
[00:23:32]  Okay.
[00:23:34]  The last question is,
[00:23:40]  You have currently any opportunities in your hand?
[00:23:44]  Yeah, after this interview, I have another interview.
[00:23:48]  Till time I don't have any kind of such a joining offer or like I was selected for HCL Tech, but they did not give me a joining letter or as well as offer letter.
[00:23:59]  But they called me, you have selected in the interview and I have given assessment in interview, but still I don't have any offer letter.
[00:24:07]  Okay.
[00:24:09]  Okay.
[00:24:10]  That's fine.
[00:24:12]  So how was my feedback?
[00:24:15]  Good.
[00:24:16]  That's it.
[00:24:18]  Thank you so much.
[00:24:19]  Okay.
[00:24:20]  We'll let you know.
[00:24:22]  Yeah.
[00:24:24]  Thank you.
[00:24:25]  Thank you so much.
[00:24:34]  Hello, Rajesh Shuriara.
[00:24:38]  Yes, Sanjay.
[00:24:39]  Hi, hi.
[00:24:41]  So, itan ki kocham meru shortlist jayate.
[00:24:45]  Okay, next round ki shortlist jayate.
[00:24:47]  Next round ki shortlist jayate.
[00:24:49]  Okay.
[00:24:50]  And then next round interview naan rahega, vare man offer sunyaal emboxa tanakodde.
[00:24:56]  I mean, tanakodde.
[00:26:02]  Okay, okay, sure.
[00:26:03]  Okay, okay.
[00:28:07]  But, ekku anaku appearance back end lo ganapistana, I am shortlisting those people.
[00:28:12]  Okay, cool.
[00:28:13]  Marvi front end leku naate kustom marli.
[00:28:15]  Aha, aadi.
[00:28:16]  Minimal.
[00:28:16]  Aadla ki chus ta.
[00:28:17]  Kachta aadla ki chus ta.
[00:28:19]  Basic knowledge on front end, strong knowledge on back end unneta to use ta.
[00:28:23]  Strong knowledge on back end, strong knowledge on ML.
[00:28:46]  Sorry?
[00:28:47]  Okay.
[00:28:49]  Team member.
[00:28:49]  Sorry.