How CT-Shift works?
CT-Shift automates and accelerates SAS workload migration to Databricks through a multi-stage process powered by GenAI. This streamlined approach minimizes manual intervention and ensures a smooth transition.
Data Ingestion
SAS datasets from sources like on-premises flat files, landing layers, SharePoint, or LAN servers are ingested into Azure Data Lake Storage (ADLS).
SAS File Parsing
SAS scripts are parsed to identify the token size for all SAS artifacts (PROC, MACRO, DATA). This involves generating SAS URLs and processing files into small chunks as per token limit.
Resource Interaction and Response Handling
Chunks are forwarded to the Embedding model to create embeddings. Databricks vector search to find out relevant functions from Knowledge base. Fetched details are then combined and fed to Databricks DBRX for generating the final response.
Code Translation
Azure OpenAI and Databricks DBRX are used to translate SAS code to PySpark/SparkSQL.
Code Generation
Translated PySpark/SparkSQL codes are received as from the API. Final PySpark file is generated by combining all the codes.
Delta Lake
Code are available on Delta Lake.
CT Shift manages
Technology Adoption and Change Management, Data Modeling, Data Analytics, Prompt Engineering and Unity Catalog.
