YOE: 13+
Current: Technical architect Big data
Referred by Friend:

There are no materials available to prepare for DE so I focused on the materials shared by FB. Have signed NDA so ony sharing on high level.
Round1: Design an app for video streaming.
product sense: First think of metrics you would like to monitor.
next build datamodel based on this.
sql: Based on your datamodel they will ask multiple sql qstns like what is avg viwer per region, avg viewing time , users who viewed only certain content.Etc

Round2: Ecommerce step design. What metrics should u look at.
SQL: Write sql to find avg time in each step.
Python: Assume u now have streaming data. Write code to find avg time per step after each record comes in and how will u update that wth constant memory.
visualization: Create some visualizations to define the metrcis which u mentioned earlier.

Round3: Given a chart, what do u understand from it. What could possible reasons of the patterns in the chart.
SQL: Given two tables write some complex relations between them using joins, case within counts, also create etl to load from one table to another.
Python: A problem to simulate group by in SQL over multiple attributes.

Ownership: Bla bla bla. Conflict, why facebook, best projct.etc etc

I tried very hard during my prep period to reach out to ppl for onsite qstns but noone actually helped for DE. So the entire prep was based on past exp and the material. I got rejected recently for Uber in last round and those interview qstns helped in the learning.

Prep: Leetcode 200 in 2months (Though the python qstns were easy standard)
Datawarehouse: Kimball
White papers of FAANGS
Quaterly report of FAANGS.
ETL, SQL: No extra prep, just revised stuff I already know.