• Data-ingest

  This consists of a high-performant MQTT broker and a highly-scalable, multi-cluster Kafka messaging queue which together can ingest and process data at a very high speed.

• Data-transform-&-storage

  This offering is built on top of Kafka KSQLdb and Apache Spark with real-time data transformation and the ability to integrate real-time machine-learning predictions for anomaly detection and other predictive maintenance tasks. The platform offers out-of-the-box connectors to store data on almost any database, data warehouse, or even into S3-storage compatible data lakes with efficient storage formats like PARQUET and AVRO.

• Data-query

  This offering consists of multiple custom-built connectors to various SQL and pseudo-SQL databases and storages along with highly-scalable PRESTO (a.k.a prestodb)

Industrial AI Vision

• Pre-trained ML libraries

  I. Object-detection built for custom use-cases such as pallet-detection in warehouses or personal protection equipment (PPE) detection in construction sites or glove-detection in process industries etc II. Optical Character Recognition & number plate detection III. Face Recognition & identification

• Vision ML Pipeline

  This enables capturing images for ML training and versioning and deployment of ML models
• for inference