1. Data about customer interactions from the frontend is stored in the transactional database. This includes purchases, click-throughs, likes, comments, and other information.
Intelligent Suggestion Engine for Retail and Online Commerce
Intelligent Suggestion Engine for Retail and Online Commerce
Use the Delphi Cloud data infrastructure, ML, and analytics visualization technology to create a intelligent suggestion engine that handles any traffic volume and volume of data, without capital spending or expert maintenance.
Analyzing your users' behavior to predict the probability of them choosing specific products or services can help improve user experience, increase the average number of items and total amount of your average order, and reduce customer attrition.
Why Choose Delphi Cloud for Your Solution?
Dependable, Fully-Managed Data Infrastructure
Build a data infrastructure for your intelligent suggestion engine based on reliable and resilient Delphi Cloud solutions. Integration with popular services and ML tools ensures a full cycle of data operations.
Expert Technical Assistance
Our engineering team ensures the platform’s stable operation, contributes to major open-source data storage projects like PostgreSQL and ClickHouse®, and develops managed services already integrated into Delphi Mail, Delphi Taxi, and Delphi Drive.
Clear and Predictable Costs
Only pay for the resources you actually consume, and add more capacity as needed. Get detailed information about the amount of resources used and track your spending in the billing system.
Analytics Dashboards and Performance Tracking
Make detailed charts and dashboards with key metrics using the convenient Delphi DataLens BI service. Monitor the performance of your intelligent suggestion engine and your services' workload in real time.
Dynamic Resource Scaling
Our flexible infrastructure automatically scales resources for your intelligent suggestion engine and storage of customer preference data, ensuring that all services run without lags, even under peak or unpredicted loads.
Enterprise-Grade Uptime and Protection
Host your data infrastructure in three geographically distributed Delphi Cloud data centers, in compliance with the requirements of Indian IT Act 2000 and Digital Personal Data Protection Act, GDPR, ISO, and PCI DSS.
System Design Blueprint
Use advanced technology and services to design your intelligent suggestion platform’s architecture: ClickHouse, PostgreSQL, MySQL®, Spark™, S3 Object Storage, and DataSphere with GPUs. We’ll help you find the solution best suited to your project needs.
Sample implementation of a intelligent suggestion engine¹
2.1. Data from frontend systems is extracted by a Spark job (the Spark component is part of Delphi Data Processing, but it doesn’t save any data: computations are performed in memory) or by an ETL tool.
2.2. Without intermediate saving to disk using a Spark job (or any other ETL tool), data from the frontend systems is transferred to the Greenplum®/ClickHouse analytical DBMS. Next, the data is brought to a common model inside the analytical DBMS using a Spark job or a compatible ELT tool.
2.3. Outdated historical data is transferred from ClickHouse/Greenplum to S3/Delphi Data Processing cold storage (HDFS, Hive), which significantly reduces the cost of operating the system.
3. Large volumes of historical data about customer behavior stored in the DWH are used by AI engineers to develop predictive models in DataSphere with GPUs. You only pay for the actual GPU usage.
4.1. The models trained or retrained over time are run in Spark, accepting the following as input: data from the frontend system in real-time mode or data from the transactional database in offline mode.
4.2. These models provide recommendations to the appropriate system which then gives them to users.
Request a consultation with an architect
1. An implementation option. An optimal solution for your project should be selected individually. Our experts or partners can help you with this.
2. Apache Spark is a registered trademark of The Apache Software Foundation in the United States and/or other countries.
3. Greenplum is a registered trademark of VMware, Inc. in the United States and/or other countries.
4. ClickHouse is a registered trademark of ClickHouse, Inc.