SaaS platform with Machine Learning capabilities to track health of industrial equipment, proactively discover anomalies and prevent downtime.
IoT streaming, Real-time data integration, Machine Learning
Imantics is a cloud-based predictive analytics platform for semiconductor device fabrication, using IoT to provide real-time IoT to provide. Imantics customers are typically large-scale fabs whose factories with multiple manufacturing automation solutions. The Imantics cloud platform can be deployed on-premise, as SaaS, or as a hybrid model on the cloud. Imantics customers cut time to market by focusing on continuously improved yields and process uptime.
The Challenge. Semiconductor fabs rely on continuous end-to-end manufacturing flows, carefully timed. Interruptions are costly. A five-minute machine malfunction can halt the whole pipeline, costing millions and destroying work-in-process material. The explosive growth in network-connected things has intensified competition and pressure on yields. Machine operators who can intervene with timely precise preventive measures can grow savings by orders of magnitude.
The Solution. Imantics has a cloud analytics platform that handles a broad variety of data types and inputs streams. It had to work hand-in-glove with complex data algorithms. CloudGeometry architects chose Kinesis stream processing to simplify ingestion of IoT device payloads. Certain cases required re-preprocessing with Spark in-memory transformations; these were launched using containers orchestrated by Kubernetes clusters. In other cases, batch processing via EMR could transform log data stored in S3, and warehouse the output in AWS Redshift.
AWS Sagemaker ML modeling framework provided a model-driven baseline; it allowed operators to compare historic performance and be certain that successive runs produced continuous improvement.
As Imantics added manufacturing customers with different IoT dat source, CloudGeometry’s CI/CD Solution simplified updates, improving the onboarding process and accelerating time to value in yield process improvements.
The Benefits. A closed-loop development and deployment process made integrating business as usual rather than the exception. Because they could more easily incorporate changes in data streams to both batch analytics and iteration of machine learning models, Imantics customers could drive continuous improvement in yield even as they ramped up production volumes.
Cleaner ML-ready data, updated continuously
Stream machine-originated data into the platform, define data transformation and processing rules
Closed-loop development for IIoT
Fully automated DevOps pipeline; configure & deploy new device payloads in minutes, not weeks
Fast, frequent, and repeatable
Integrated CI/CD automation for frictionless, agile business innovation