Smart Deployment Manager

An ML-powered One-click Cloud deployment manager. The user just needs to pass a simple JSON object with the details of his application along with the required cloud vendor’s credentials.His application will be deployed on his selected cloud like Google Cloud Platform along with the best-suited infrastructure implementations automatically.

 

  • Purpose: Make the long and difficult process of application deployment much easier, faster and efficient.
  • Connection to Business: It’s a complete solution for application deployment, so application owners and developers can subscribe to use this service against charges.
  • Revenue: As it’s machine learning powered, so the users don’t need to hire someone to make the deployments, which is highly cost saving.
  • We achieved the training of this very difficult machine-learning model, even though it was a classification problem but the data structure makes it much difficult.
  • Successfully deployed and scaled hundreds of real-world applications.

 

  • Challenges:  The very big challenge was the data because we were needed to train the model to investigate the best-suited combination of cloud resources on the base of various properties. I overcome this issue by contacting with various universities and request them to ask their students to use our application to deploy their final year projects and then we monitor them deeply.

 

  • Learning: 
  1. I got the deep experience to build applications in the form of micro-services.
  2. Learned various APIs from GCP & AWS.
  3. Learned how to create the data from very scratch and passed it through various steps like normalization and pre-processing.

 

  • Contributions:
  1. I have designed the complete architecture of this project.
  2. Implemented various APIS from GCP & AWS.

© 2020 Pythonist.org | All rights reserved | Design by W3layouts.