Guntur Cahyono
Student, Universitas Gadjah Mada, Indonesia
I’m writing this based on the recent Ilum introduction online meeting and a small experiment I’m testing using Ilum on Minikube. Getting introduced by my lecturer Mr. Widyawan about the existence of this technology drives my curiosity to explore. Main idea that I’ve introduced and I’ve tested is that the cut-off between the initialization of Spark-job to run by interactive Spark session is really great.
Moreover, in this latest release (that I’ve installed on Minikube) basically everything that is unexpectedly covered such as the lineage dashboard, metrics on the running job, and also how the data from running SparkUI gets aggregated into an easily accessible dashboard. I’m pretty interested in the case that in the introduction demo that the capability to execute LLM models are also included and the way to present between the payload to result-respond also helpful. As of currently we run this on educational purpose while my background also came from industry, Ilum drives a lot of potential use-case to be tested on big data especially since on last introduction meeting the Ilum team gonna integrate Apache Doris as additional for analytical purpose. Surely I will hear further and more about this analytical usage and simplicity.