Scribd
Upload
33 views

Cloud Architecture Recommendation Using LLM Model (2)

Uploaded by

sajijayden
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
33 views

Cloud Architecture Recommendation Using LLM Model (2)

Uploaded by

sajijayden
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Cloud Architecture Recommendation Using LLM Model

Overview

The goal of this project is to build a recommendation system that suggests cloud architecture to clients based on their product requirement and their preferences.
The recommendation system will be built by constructing an LLM Based OpenAI Based word embedding.

Dataset
There is no dataset which is openly available to do this project. We need to create the dataset which is text data that need to be trained upon. The suggestion am put
forwarding are the follows

AWS cloud architecture books


AWS Documentation
Cloud Blogs
Cloud architecture research papers and studies (Open Source)

Methodology
The recommendation system will be built using the following methodology:

1. Data Gathering: A large dataset of text documents related to cloud architecture design will be gathered either from open-source datasets or by scraping the
web.
2. Preprocessing: The text data will be preprocessed to clean and normalize the text using techniques like tokenization, stemming, and lemmatization.
3. Training: Training the data using OpenAI word embedding, Langchain and GPT 3.5.
4. Prompt Engineering: Prompt engineering back end prompts to give the desired out in the language context we need
5. Fine Tuning: Based on the output given, fine tuning the hyper-parameters, troubleshooting the dataset for corrections, Improving accuracy.
6. Graph Creation: Convert the current input text data into a csv table, which then converted to a graph.

Tools
The recommendation system will be built using Python and the following libraries:

Pandas: For data cleaning and preprocessing.


GPT 3.5: For training and creation of a generative AI based on the data we are given.
Langchain: Connect GPT 3.5 to the data we have prepared for training.
NetworkX or Graph-tool: For coming up with the user diagram.
Surprise: For building collaborative filtering models.
Flask: For building a web application to demonstrate the recommendation system.
Angular: For the frontend development purposes.

Deliverables
The project will deliver the following:

1. A natural language processing (NLP) application that generates cloud architecture designs from text input using LangChain. Which will be then converted to a
graphical diagram.

2. An LLM model trained on a large dataset of text documents related to cloud architecture design that generates cost-effective, scalable, and low-latency
designs.

Conclusion

Building an LLM Model that generates cloud architecture designs from text input using LangChain is a challenging but potentially valuable project. Implementing the
project comes with several risk which needs to analyzed and thoroughly researched

You might also like