Data Engineering at Comcast

The employer bargain at Comcast sits in familiar large-enterprise territory: relative job stability, a mature data org with real infrastructure, and pay in line with the other companies we track. What you give up is pace. Blind sentiment runs mixed and happiness is stressed and trending down, which aligns with what you hear from engineers inside large legacy telecoms: the data problems are genuinely interesting but the org layers and approval cycles slow execution. The tension worth naming is that Comcast operates a content business (Peacock, NBCUniversal) inside a network infrastructure company, and data engineering priorities can shift based on which internal client has budget that quarter. Engineers who want to ship quickly and see direct product impact will find that harder here than the interesting-problem surface would suggest.

3 tracked layoffs in the past 12 months, the most recent in Jun 2026 alongside 3 executive departures over the same window, which is a meaningful signal at a company of this size and warrants attention. Layoff risk over the next 30 days reads moderate. Hiring volume is modest: 5 open data engineering roles across 2 cities, concentrated in Chicago. That combination of recent workforce reductions, leadership turnover, and a thin open-role count suggests Comcast is in a consolidation phase rather than an expansion one. Someone joining now should expect to inherit existing pipelines and focus on reliability and cost efficiency rather than building net-new capability. The trajectory over the next 12 months looks like continued pressure to do more with a leaner team, which can accelerate growth for the right engineer but is a real constraint on greenfield work.

Comcast's core data problem is scale across a consumer base measured in tens of millions of cable, internet, and streaming subscribers, with behavioral signals arriving from set-top boxes, mobile apps, Peacock viewing sessions, and network telemetry simultaneously. Data engineers here are building pipelines that connect ad-targeting, content recommendation, and network operations domains, often under regulatory constraints around subscriber privacy that add surface area to every schema decision. The visible stack centers on AWS, Azure and Docker with Java, Python and Scala as the working languages, which points toward a cloud-hybrid architecture where on-premise network data meets cloud-native processing. That combination means the day-to-day likely involves managing cross-environment data movement and latency tradeoffs, rather than a clean single-cloud setup, and ownership tends to span from ingestion through to the warehouse layer.