To use R scripts with Flyte, you can utilize raw container tasks. This approach allows you to run R jobs as part of your Flyte workflows. Here's a summary of the steps and considerations based on the discussions and information available: 1. Raw Container Tasks: You can run R scripts using raw container tasks in Flyte. This method involves creating a Docker container that has R and the necessary packages installed. You then specify this container in your task definition to execute the R script. Kevin Su confirmed the use of raw container tasks for R jobs in a discussion on the Flyte forum. 2. Data Passing: If your workflow involves passing complex data types (like dataframes or CSV files) between Python and R tasks, you can serialize the data into a format that both languages can understand. For example, you can pass dataframes as CSV or Parquet files. This approach was discussed by Cody Scandore, who mentioned the need for passing dataframes or CSV files between Python and R tasks. Ketan (kumare3) suggested using R2py or reticulate for invoking R scripts or passing the data as Parquet or CSV files. 3. R SDK Consideration: There was a discussion about the potential use of an R SDK with Flyte to facilitate R tasks. This indicates an interest in making R integration more seamless in the future. 4. Examples and Documentation: Currently, there may not be specific examples in the Flyte documentation for using Parquet files in R tasks. However, the community has expressed interest in creating such examples, which suggests that more resources may become available. In summary, to use R scripts with Flyte, you should use raw container tasks, consider the data passing mechanism (possibly through intermediate files like CSV or Parquet), and stay tuned for potential developments regarding R integration with Flyte.