Abhishek effectively bridges the gap between abstract vector theory and practical engineering, making complex RAG concepts accessible without sacrificing technical depth. This is a high-signal resource for anyone looking to build AI applications that actually understand context rather than just matching keywords.
Deep Dive
Prerequisite Knowledge
- No data available.
Where to go next
- No data available.
Deep Dive
Complete FREE Real-time AI Projects + Workshops for BeginnersAdded:
Hello everyone, my name is Abhishing and welcome back to my channel. What if I tell you you can learn AI for free practically and by building realtime projects. Yes, let me get straight to the point. In today's world, we all want to learn AI. It can be MCP, it can be rag, it can be AI for DevOps, AI for cloud. But one of the major challenges that we all face, most of the resources out there are purely theoretical.
But what we actually need is something real time and project focus. In today's video, I'll introduce you to a gold mine. Basically, a GitHub repository that has realtime projects, notes, and even workshops.
Oracle has recently open sourced this GitHub repo. In today's video, I'll walk you through the repo. I'll walk you through the realtime projects and most importantly I'll demonstrate one of the projects for you. All that you have to do is to watch this video till the end.
I'll share the link to the GitHub repository in the description and also in the pinned comment. Just start the repository and keep learning AI. Let's get started.
So this is a repository Oracle AI developer hub. Again the link is in the description. just start the repository because they're adding a lot of new projects. So what's in the repository?
As I told you, they have real world applications.
Some of the applications are listed here, but I'll walk you through all the applications. Basically, using these applications, you can understand how real world AI projects are built, how are they deployed, and how you can invoke or inference these AI applications.
Then there are also notes. Let's say you want to understand system design or architecture diagrams. The note section can help you. Then most importantly they also have workshops.
Workshops can help you understand end to end use cases. Let's say you want to understand end to end rag use cases. You can go through the rag workshop. Or maybe you want to understand from racks to agents everything. You can go through this workshop. I've recently gone through this one. agent memory even you can follow this workshop but Abishek where do I get started okay there is a lot of information in the repository the best place to get started head to the applications and start with Oracle database vector search because vector search is beginner friendly concept and in fact you can implement this entire project in less than 30 minutes. The good part is that for every application they have detailed blog. I mean wherever it is required they have provided a detailed blog that they have implemented. I mean they have explained the architecture of it. They have explained the source code how different components of these project interact with each other. And then in the readme file they also provided the operation information. What is operation information? How to run the project? how to deploy the project, how to use AI within the project and how to invoke the API of the project. All of that which DevOps and cloud instance deal with is also provided in the readme file. The same is with every application. Let's say you want to understand rag for rag.
They don't have a blog but all the information is provided in the readmi file and once the information is good they have provided the operation information. So for all the 10 applications you can find the similar information. Once you're done with the applications then you can head to the workshops.
Cool. So let's go to Oracle database vector search. Let me first walk you through and explain what exactly is this project. So I'll take the whiteboard and typically explain you what is in this block.
So we are going to learn vector search to a pet store application. But Abishek what exactly is vector search or what exactly is a vector database? Simple.
If you look at a traditional database imagine you are working for an e-commerce platform.
The list of products information right in a typical e-commerce platform there are products list of products information let's say is stored within the database.
Now imagine these products some of these products are related to dog chains.
So in the e-commerce platform you have let's say 100 chain products.
Now a user goes to the platform and let's say user search for dog belt.
Okay, maybe a common user search for dog belt in a traditional database exact search is performed or exact match is performed. So although the e-commerce product or the platform they have 100 dog chains because user search for dog belt the exact match is not found. So user would see no results found which is a huge loss for the platform because what user is looking for is actually available on the platform. This is where vector search and vector DB is help.
Because when it comes to vector databases, exact search is not performed. Semantic search or similarity search is performed.
So what vector databases do or what engineers do? So they take the product information using open AI or using other models. They convert the product information into embeddings.
and they store the embedding information onto the vector databases.
Now things become very very simple. If you take the same use case, let's say user search for dog belt, all the dogchain product information is stored as embeddings into the vector databases.
Vector databases perform similarity search and it shows the users all the dog build information. So basically the similarity search is a key factor in vector databases and to achieve that engineers take the product information and then convert the information into embeddings.
But Abishek how are we going to learn this in today's tutorial very simple step one we will install Oracle database which is a very popular vector database and the good thing is that Oracle database is also free. You can just run the database as a docker container or maybe using podman.
Second, once you have the Oracle database, you need the source code, maybe a Python application or a Java application that takes the product information and use the open a embeddings and convert the product information into embeddings.
The good part all of this is done by the application.
If you go through the blog, they have also explained how this embedding is performed and how the product information is stored onto the database.
I'll not go into that detail because it is going to take a lot of time. But the good part is that the source code is already available. So step three, we will just run the application.
Now when we run the application it connects to the Oracle DB and database is all set up along with the password and everything.
Only thing that we also have to do we need to set up OpenAI API key.
So this is the only thing that you have to set up. Additionally for this project you cannot use Olama.
But let's say this is a restriction for you. Let's say you cannot use OpenAI API key for any reason. Don't worry. There are a lot of other projects where you can actually use Olama as well. For example, just go to tanstack shoe store within this. You can use Olama or you can use claude, you can use GBT4, anything. This application supports all of the models. Perfect. So this is all we have to do. Step one, we have to install the Oracle database.
Then we have to clone the repository. We have already start the repository. So just go there, clone the repo.
Before that just store the database password within an environment variable and store open API key within an environment variable. Finally run the application. Now all of this information, how do I know all this information? because I've already executed this application.
So this is the same information that I'm trying to explain you. See, download the database, then store the database password, open a API key, and finally run the application. Let's do that. So first we need to clone the applica repository.
Let me pull my terminal. So you can see here I am already within Oracle database vector search. So I've cloned the repo Oracle AI developer hub and I move to the apps folder and I have Oracle database vector search. So these are all the files.
Now I need to install the Oracle database. Abishek. How do you do that?
Just run this docker command. In fact the docker command is also provided here. If you just scroll down, this is the docker command. Aisha cannot run docker on my machine for any reason.
Don't worry, you can also go to Oracle.
You can create an account with Oracle and within the Oracle cloud, you can create Oracle 26 AI database which is part of the free trial. So just go to the Oracle cloud and you can also implement it.
because I'm able to run docker. I have used the docker command and I have the Oracle database running. If you observe carefully, this is the password that I have provided. Oracle password is my password. So that's why I have exported the password as well. Similarly, I have an open AAI account. I have also exported the OpenAI secret key.
Definitely, I cannot show that in the video. So I just ran this command export open a API key followed by my personal API key and then I will just run this command now mvn spring boot run. But Abishek where are the products? You can see that as soon as I run MVN spring boot run you will see the list of products and how they are stored as embeddings within the vector database in our case the Oracle database. See these are the products Labrador back bar control chiefs. So this is an embedding. So these are stored in the vector database. Similarly heavy duty rope silent laser pointer different products. Hairbrush for long hair cats fish tank for small fish. Now in a regular database if I only search for fish tank for small fish this product information is returned or maybe I have to search one of the strings in this pa or strings in this sentence maybe I have to search for fish and all the fish information is resulted right all the fish information is returned now let's see what happens I'm running this application you can see this is running it's running on my local host I will take insomnia.
If you want, you can use the curl command as well or you can use postman.
It's up to you. And just run this command in the documentation. You also have the curl command.
So this is the one.
Just copy and paste it in insomnia or your browser curl anything.
Let me remove curl and execute it. So what I'm running treats for small loud dots.
Let me show you this side. So this is the request and this is the response.
You can see I did not search anything that's here. I search for treats for loud dogs. Or let's change it. Okay.
What I'm going to do is I'm going to make it easy. So I'll change this query here and let me say dog belt. Definitely there is nothing here with dog belt. If you observe these things here there is nothing related to dog belt. I mean there is something related to dog belt.
For example this one here heavyduty rope but it doesn't say dog belt. Let's see if this information is resulted. Okay, I have to get this information because I'm searching for dog belt and this is heavy duty rope for large dog breeds.
Send.
There you go. Heavy duty rope for large dog breeds. And if you observe carefully, none of the other information is resulted because similarity search is performed here. I will also change it.
Let's say I will search for something like fishbone.
Let me run the it resulted fish tank for small fish. If you observe again there is nothing called fish ball but I got the information fish tank for small fish because it performed the similarity search. It understood okay you don't have fish ball but you have something like fish tank. So that's why I'm returning the result here. Perfect. So this is how vector search is actually performed and in fact we implemented this entire project in less than 30 minutes. You might take more time if you want to go through the complete blog but only if you're interested in the internal implementation. If you are a devops or cloud engineer maybe you can focus on the operational information.
But if you come from a development background there is lot of information here. For example, how you can use Gwv VM instead of JVM. That is how to start the Java application faster when you are in the AI world and especially when it comes to vector databases. They are very useful in the world of AI. Let's say you're building a chatbot or maybe you're using rag. In the world of rag, you need similarity information, not exact information to be retrieved from the database.
Perfect. So I hope you found this video informative. There is lot of content in the repository. So make sure you make the best use out of it. Try different projects and try different workshops.
I'm sure they are going to add more projects to the repository as well. Let me know in the comment section if you have any questions or if you want me to make video on any particular project.
See you all in the next video. Take care.
Related Videos
Agentforce NOW AMA: Build with React and Salesforce Multi-Framework
SalesforceDevs
490 viewsβ’2026-05-28
How agent o11y differs from traditional o11y β Phil Hetzel, Braintrust
aiDotEngineer
450 viewsβ’2026-05-28
WEB TECHNOLOGIES UNIT-2 | Degree 4th sem BCOM Computers web technologies unit-2 full explanationπ―β
LearnwithSahera
1K viewsβ’2026-05-29
More tests are always better? How to use AI to identify tests that bring little value
Alliance4Qualification
335 viewsβ’2026-05-29
Search Algorithms Explained in 60 Seconds! π€π¨
samarthtuliofficial
218 viewsβ’2026-06-01
People of Game of Thrones using JavaScript DOM
AltCampus
296 viewsβ’2026-05-30
Introduction to Problem Solving Part - 1 | Lecture 1 | Intermediate DSA
ascensionix
107 viewsβ’2026-05-29
So What's Odin Lang Even Good For
TechOverTea
131 viewsβ’2026-06-01











