A Data Management and Machine Learning Platform

Predera Technologies is a Big data and Machine Learning company building AI solutions for Healthcare, Finance and Retail industries. Underlying these solutions is our Data Management and Machine Learning Platform which helps in rapid development and deployment of new solutions.

The Platform helps fetch data from multiple sources and validate, cleanse, process and store for analysis and reporting. The platform provides a drag-and-drop pipeline builder using which the users can quickly build pipelines for data ingestion, ML model building or real-time scoring. The visual interface helps preview data and understand the data flow. Data quality, security, monitoring and reporting are built-in and configurable. The platform provides connectors to various data sources and destinations, processors for data transformations and feature engineering, and ML algorithms for building models. Plugin architecture makes building custom connectors and processors easy. Java and Python programming languages are supported.

This platform is not only helping Data Engineers and Data Scientists at Predera build new solutions quickly but is also being used by our Clients with the flexibility of on-premises, cloud or hybrid deployment options. Data Scientists can use the platform for both building and deploying Machine Learning Models. Existing ML Models in pickle, PMML, POJO and MOJO formats can be imported and deployed on the platform. Platform also supports integration with ML Models deployed elsewhere (such as H2O, Azure ML, etc) so that the data ingestion and scoring pipelines are at one place.

Features of the platform

  • Data Ingestion from multiple sources
  • Data Modeling, Mapping and Transformations
  • Data Processing (Batch and Stream)
  • Data Quality Management
  • Data Security and Governance
  • Data Services (e.g. Patient Data Service)
  • Monitoring and Reporting
  • Machine Learning Algorithms and Model building on big data
  • Machine Learning Model scalable deployment and real-time scoring
  • Reports and Dashboards
  • Rich UI with Role based user experience

The platform is built on open-source technologies and is highly configurable and extensible.

Architecture

  • Standalone mode or cluster mode (using Spark)
  • Container Architecture
  • Allows deploying on any cloud platform
  • Built on top of open source technologies
  • OAuth2 and role based access
  • Data security at User, Dataset and Field levels; secure transmission and storage
  • Plugin architecture enables custom connectors and processors - Java and Python programming languages are supported
  • REST APIs for information access, scoring engine and custom integrations
  • Supports embedding BI reports and dashboards

 

Figure: Platform Architecture

Benefits

  • Rapid development, testing and deployment of new solutions
  • Integrated Machine Learning Platform
  • Self-Service and role based user experience

Use Cases

The platform is being used internally at Predera to build solutions for Clients in Healthcare, Finance and Retail. Below are some of the solutions

  • Healthcare - Real-time Sepsis detection, Readmission and Length-of-stay prediction
  • Finance - Credit Decisioning System, User spending behavior segmentation
  • Retail - Customer Segmentation and Recommendation Engine

Please contact us at hello@predera.com for a live demo or to learn how the platform can expedite your company's move to artificial intelligence.