Databricks Software Engineer Interview Questions

Review this list of Databricks software engineer interview questions and answers verified by hiring managers and candidates.
  • Databricks logoAsked at Databricks 

    "High Level Architect Client v API Gateway v Object Storage v Message Queue v Worker v Database Client should can document with a web site or directly with API services. API Gateway should be used for upload document,get document info and state. Object storage should be used for original document and send event to Message Queue for starting. Message Queue is neccessary because there are millions of document should be process each time. Worker can get text from document with OCR. Database shoul"

    Berk C. - "High Level Architect Client v API Gateway v Object Storage v Message Queue v Worker v Database Client should can document with a web site or directly with API services. API Gateway should be used for upload document,get document info and state. Object storage should be used for original document and send event to Message Queue for starting. Message Queue is neccessary because there are millions of document should be process each time. Worker can get text from document with OCR. Database shoul"See full answer

    Software Engineer
    System Design
    +2 more
  • Databricks logoAsked at Databricks 

    "Constraints: 4-direction moves; no mode switching (pick exactly one of {1=bicycle, 2=bike, 3=car, 4=bus} for the full trip). Per-mode search: If a mode’s per-step time/cost are uniform, run BFS on allowed cells. Then totaltime = steps × timeperstep, tie-break by steps × costper_step. If time/cost vary by cell (given matrices), run Dijkstra per mode minimizing (totaltime, totalcost) lexicographically. Maintain the best ⟨time, cost⟩ per cell; relax when the new pair is strictly better. S"

    Rahul J. - "Constraints: 4-direction moves; no mode switching (pick exactly one of {1=bicycle, 2=bike, 3=car, 4=bus} for the full trip). Per-mode search: If a mode’s per-step time/cost are uniform, run BFS on allowed cells. Then totaltime = steps × timeperstep, tie-break by steps × costper_step. If time/cost vary by cell (given matrices), run Dijkstra per mode minimizing (totaltime, totalcost) lexicographically. Maintain the best ⟨time, cost⟩ per cell; relax when the new pair is strictly better. S"See full answer

    Software Engineer
    Coding
    +1 more
  • Databricks logoAsked at Databricks 
    Software Engineer
    System Design
    +1 more
  • Databricks logoAsked at Databricks 

    "One Accomplishment I'm most proud of is that I graduated from Schaumburg High School In May of 2021 and I was able to get up the stage and collect my diploma. This was a HUGE Impact in regards of passing all of my classes and earning all of my credits in order to be apart of the NOW Arena graduation."

    Amparo L. - "One Accomplishment I'm most proud of is that I graduated from Schaumburg High School In May of 2021 and I was able to get up the stage and collect my diploma. This was a HUGE Impact in regards of passing all of my classes and earning all of my credits in order to be apart of the NOW Arena graduation."See full answer

    Software Engineer
    Behavioral
    +1 more
  • 🧠 Want an expert answer to a question? Saving questions lets us know what content to make next.

Showing 1-4 of 4