Automated Model Inferencing Pipeline for Real-Time Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data pipeline monitoring systems require manual, hardware-intensive efforts to build model-specific pipelines for each desired metric, limiting the number of individuals who can develop and deploy models due to technical complexity and high overhead.
Innovation Solution
A system and method for automatically generating and deploying a model inferencing pipeline using a training dataset from a cross-customer data pipeline to extract features from a customer-specific data pipeline, reducing the need for manual construction and increasing accessibility for model deployment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual model-specific pipeline construction is used, then pipeline customization for each model is achieved, but hardware resources and time consumption increase significantly
Solution Approach 1:
The system performs preliminary action by pre-extracting and storing feature pipelines from training data sets before actual model deployment. When a new model needs to be deployed, the system retrieves pre-extracted features from storage rather than constructing pipelines manually, significantly reducing hardware resources and time consumption while maintaining customization capabilities.
2Adaptability or versatility
If manual pipeline construction is required, then model-specific feature extraction is achieved, but the number of individuals who can deploy models is limited
Solution Approach 1:
The system implements self-service by automatically extracting features from training data sets and storing them in feature stores. When models need to be deployed, the system automatically retrieves and utilizes these pre-extracted features without requiring manual pipeline construction. This automation reduces technical complexity and enables more individuals to deploy models while maintaining model-specific feature extraction capabilities.
3Measurement precision
If manual pipeline building is performed for each model, then model-specific metric calculation is achieved, but deployment time and overhead increase
Solution Approach 1:
The system performs preliminary action by pre-extracting and storing feature pipelines from training data sets before model deployment. When a new model is deployed, the system retrieves pre-extracted features from storage, significantly reducing deployment time while maintaining metric calculation accuracy through the use of pre-processed, high-quality feature data.
Data Source
AI summary
Systems and methods of real-time modeling pipeline inferencing are disclosed. At least one model configured to calculate at least one metric from one or more features is deployed. A model inferencing pipeline configured to extract the one or more features from a customer-specific data pipeline is implemented for the at least one mode. The model inferencing pipeline is generated using a training data set extracted from a cross-customer data pipeline. The at least one metric is calculated using the one or more features extracted from the customer-specific data pipeline.