Data Engineering at Snap

Snap pays above other Technology companies, and the comp numbers at senior and staff levels are real. What you trade for that is a workplace under visible strain. The 3.4 Glassdoor rating is toward the bottom of the pack, Blind sentiment is mixed, and the internal happiness picture is stressed and trending down. The tension is specific: Snap has been public about the pressure to grow revenue on a platform whose user base stopped compounding at the rate that justified its earlier valuation. Engineers inside that pressure tend to describe shifting priorities and reorgs rather than a stable platform mandate. Pay is genuine compensation for that instability, and for some engineers it's the right trade; for others, the organizational churn erodes the kind of deep technical ownership that makes a DE role satisfying long-term.

Snap has had 4 tracked layoffs in the past 12 months, the most recent in Jun 2026, alongside 5 executive departures over the same window. That combination signals an organization still working out its cost structure, and the low 30-day layoff risk reflects ongoing uncertainty rather than a post-restructuring calm. Hiring is still moving: 31 open data engineering roles across 15 cities, with Chicago as the top market. For someone joining now, the next 12 months will likely bring continued cost discipline and possible team consolidation, but the open role count suggests the data org is not in a freeze. The realistic frame is that Snap is a company investing in DE capacity while managing headcount pressure elsewhere; a new hire should expect to land in a team that is productive but where the broader environment stays unsettled.

Snap's core data problem is real-time at scale: a camera-first social platform where billions of daily ephemeral messages, Stories, Map activity, and ad impressions need to be ingested, attributed, and made queryable fast enough to power both the ad auction and the product experience. The stack on current listings centers on AWS, S3 and Spark, with Python, SQL and PySpark as the working languages, which points to a batch-and-streaming shop where engineers own end-to-end pipeline work from ingestion through warehouse delivery. Snap's ad business requires tight latency on impression and conversion data, so pipelines here carry real SLA weight. The camera and AR surface adds a domain most data engineers don't see elsewhere: sensor-adjacent event streams with irregular shapes that need careful modeling before they're analytically useful. That mix makes the job broader than a typical ad-tech DE role.