Company attributes
Other attributes
KenSci offers an AI-powered, predictive analytics platform for healthcare providers, enabling automatic identification of clinical, operational and financial risks.
The company was founded in December 2015 by Ankur Teredsai, David Hazel and Samir Majure in Seattle, Washington, United States. KenSci was acquired by Providence-based healthcare AI company Tegria on June 24, 2021.
KenSci's machine learning-powered risk prediction platform helps healthcare providers and payers intervene early by identifying clinical, financial and operational risks in healthcare.
KenSci's platform is engineered to ingest, transform and integrate healthcare data across clinical, claims, and patient-generated sources. Its machine learning platform integrates into existing workflows, allowing health systems to better identify utilization, variation and improve hospital operations through a library of pre-built models and modular solutions.
KenSci’s AI Platform for Digital Health is designed with scalable, enterprise-ready data architecture that automates the ingestion, data preparation, processing and transformation of data into business intelligence (BI) and artificial intelligence (AI) ready formats within the Azure cloud.
Customers can monitor data quality across various dimensions such as completeness, semantic and syntactic correctness, morphological accuracy and other data quality factors.
The service allows quick, repeatable data ingestion from EMR, claims, devices and other public and custom data sources into an Azure-based data lake, offering migration from on-premise or cloud assets.
Pre-built data connectors integrate data into industry-standard schema and data formats with connections to FHIR, Common Data Model, Synapse, and Databricks for downstream analytics applications.
KenSci’s AI Platform comes with an integrated analytics development platform that enables in-house analytics on top of data pipelines, giving clients access to a drag-and-drop report builder. The analytics team doesn't have to source data and prep the data to create new reports, because the data pipeline is managed for quality and application readiness.
The web-based analytics portal uses a PowerBI-based report building interface and offers access to over 100 health features for creating new reports and dashboards. Auto-generated KPIs and system-wide metrics enable out-of-the-box reports and dashboards that can identify ROI opportunities based on system-level insights.
Customers can search for data via a Q&A query interface across all data fields and sources, and utilize the platform's cognitive search interface.
KenSci's platform offers multiple tools which enable AI development, including a feature library, training datasets, phenotype engine, "bring-your-own model", pre-built AI model, model lifecycle management, and streaming and branched machine learning pipelines.
KenSci's feature library holds hundreds of clinically validated healthcare attributes. These attributes are auto-generated by underlying data pipelines, providing easy-to-develop and use tags for AI and ML model usage.
AI platform allows for training datasets that reference the ingested data pipeline as well as public data sets. Training data sets are available in the AI model development workspace and allows sharing, monitoring and tracking during model development and testing.
The platform's clinically and research-validated phenotypes enable the segmentation of underlying data into common use-case applications. These phenotypes accelerate variation analysis, agile experimentation, and usage in model development and testing.
KenSci’s AI model engine allows stateless model scoring, enabling models developed, hosted and trained on other data sets to be scored on KenSci's managed data pipeline. This helps expedite in-house projects and experiments transition to production.
KenSci’s data pipeline aggregates streaming and batch data into an ML pipeline for model development, testing and deployment. This enables a wide variety of real-time data use-cases, with support to HL7 and IOMT real-time data.
The AI model management platform allows for model development, testing, deployment, monitoring and maintenance across the lifecycle of AI and ML models. This enables data science teams to bring AI-based insights into production workflows.
d
