Sony · Primly Community

Sony data engineer interview, pipelines and SQL, full breakdown of what they tested

ds_dmitri · 6 replies

Went through the Sony data engineering loop in early 2026. Role was on the analytics infrastructure team, Sony Electronics side (not SIE/PlayStation). Here's what the whole thing looked like.

Recruiter screen: 30 min, mostly background and comp alignment. They asked about my experience with cloud data warehouses and Kafka before moving me forward. Those two things seem to be a real signal for them.

Technical phone screen: 45 min. SQL-heavy. Two questions: A multi-table join query with window functions. They wanted me to explain how the query plan would behave at large scale. A short data modeling problem. Given a set of product attributes, how would you schema this in a star schema vs a flat table, and when would you choose each.

I've done a lot of DE interviews and this was in the top half for SQL depth. Most companies just test if you know GROUP BY. Sony actually cared about understanding.

Onsite (4 rounds):

Round 1 (coding): Python pipeline task. Given a messy JSON feed, parse, transform, and write to a target schema. Focus on handling malformed records gracefully. Very practical.

Round 2 (SQL deep dive): More complex analytical queries. Running totals, cohort-style analysis, handling late-arriving data in a streaming context. One question explicitly mentioned Kafka offsets.

Round 3 (architecture): Design an event ingestion pipeline for IoT sensor data from consumer electronics devices. Kafka to Spark to Snowflake was my answer. We talked about schema evolution, backpressure, exactly-once semantics. Good 60-minute conversation.

Round 4 (behavioral): Standard stuff. Collaboration, handling ambiguity, time you disagreed with a technical direction.

Stack they mentioned being on: Kafka, Spark, Airflow, Snowflake, some dbt. GCP as the cloud provider for this team.

Compensation for senior DE at this level in 2026 was around $160k-$190k base with bonus. Not top-of-market but also not insulting.

6 replies

analyst_ana

The window function question being about query plan performance is next level for a typical company. Most places just want you to write the query. Did they actually expect you to explain execution in detail or was that a bonus?

de_derek

Bonus territory, I think, but it clearly lit up the interviewer. The question opened with 'write the query' and then he followed up with 'what would concern you about running this on a table with 2 billion rows.' So the performance angle was their add-on, not the main ask.

finance_faye

GCP, Snowflake, and Airflow is a very standard modern stack. Good sign they're not on something cursed. Did they have any analyst-facing layers or was it all engineering infra?

de_derek

They mentioned a dbt layer feeding BI tools downstream. The team was exclusively engineering focused but there were separate analytics teams consuming the data. Pretty clean separation of concerns from what I could tell.

contractor_kai

$160k-$190k base is on the lower end for senior DE in 2026 SF/Bay Area. Did you negotiate or was that a take-it-or-leave-it number?

de_derek

I negotiated. Started at $158k, ended at $172k base plus a small signing bonus. I didn't have a competing offer in hand which limited leverage. If you have one, use it. Sony does have some band flexibility.