Will deeply understanding the Photon Paper help me get a job at Databricks or other places?
What else should I study besides leetcode and generic system design?
No, because it’s just, “we implemented an executor in C++”. Your time is better spent looking at either their growing application/product offerings and building a skillset to help with that development, or starting to work on open source Spark itself and seeing what you can learn and contribute.
thanks
Of course. They ask it at the start of every interview
If you are talking about in general to join Databricks, no need for that paper. Most of the Databricks engineers do not understand that paper I believe. If you are talking about joining Photon team, you need database internal experiences. That paper is not enough.
Thanks. Can you please expand more on the "database internals" part - how far back should I go in terms of history (i.e. how much to focus on stuff pre 2010 or post or whatever date) but did you have specific things that were educational? I see this https://www.youtube.com/@CMUDatabaseGroup channel but just curious if you had specific open source projects other than Spark that you had in mind. I have lots of time to do studying / hacking on things but just want to make sure I prioritize properly.
I don’t think we ever hired anyone without database experiences into photon team. The team members were either majored in database at school, and/or industry hires worked many years in Redshift/BigQuery/Cloudera etc. Just read some paper and watch online video are unlikely to be considered. Such materials are good for your education, but do not spend too much time on it only to apply for a job.
Life