ML Ops Automation and Dashboarding
Retail & CPGBusiness Impacts
Reduced Manual Effort for Set up and Configuration
Smart Insights and Visualization
Improved Operational Efficiency
Customer Key Facts
- Location : India
- Industry : Retail & CPG
Problem Context
The customer is an Indian consumer goods company that primarily manufactures fashion accessories such as watches, jewelery and eyewear. They had been building Machine Learning models on local systems, which need to be deployed and configured manually. This entire process is time-taking and the system is not scalable.
Challenges
- Automating using Airflow by leveraging parameter store for around 100 models in production
- Building a generic template to accommodate over 100 models of different frameworks
- Rigorous testing of each model’s end-to-end pipeline in development environment before deploying in production environment
Technologies
Amazon Athena
Amazon Glue
Amazon Quicksight
AWS Lambda
AWS S3
Amazon Sagemaker
AWS Code Commit
Amazon RDS
Apache Airflow
Automating Machine Learning Model Scoring, Evaluation, and Retraining
Solution
Quantiphi is helping the customer in the orchestration of training, evaluation, and deployment of over 100 models running on Amazon SageMaker by leveraging Apache Airflow. Once each model is deployed, the Production Model Lifecycle is such that the scoring, evaluation, and retraining of the model is automated. Quantiphi is also helping the customer in the visualization of the performance of the deployed models using Amazon QuickSight and Amazon Athena.
Results
- Generalized templates for deployments with ease
- Automated Production Model lifecycle management
- Smart insights and visualization on the performance of models
