Data Engineering at AMD

AMD pays above other Technology companies, which is the clearest bright spot in the signals. The 4.0 Glassdoor rating is a little above the middle of the pack, and Blind sentiment reads mixed, which is not what you'd expect from a company riding a real AI hardware wave. neutral with a roughly flat trend suggests something is absorbing the tailwind before it reaches individual contributors: execution pressure from competing with Nvidia in an accelerator race that punishes any slip, a historically conservative culture adjusting to hypergrowth hiring, or post-acquisition integration overhead across the Xilinx footprint. You get above-market pay and hard infrastructure problems. You give up some of the cultural ease you'd find at a pure software company, and the org can feel like it's running two speeds at once.

AMD's trajectory is expansion in the AI infrastructure segment, but the organizational picture is more complicated than the stock chart. 4 executive departures in the past 12 months is worth watching; senior leadership churn at this rate during a product-cycle build often means strategy is still crystallizing at the top, which filters down as shifting priorities for data teams. no tracked layoffs in the past 12 months, and low 30-day layoff risk, so the near-term seat is stable. Hiring volume is thin right now, with 5 open data engineering roles across 2 cities, concentrated in Austin, which points to selective rather than broad expansion. If AMD is disciplined about where it adds data engineering headcount, the roles that do open are likely to carry real scope rather than being backfills.

AMD's core data problem is chipmaker scale colliding with AI-era demand. The company spans CPU, GPU, FPGA, and embedded silicon, which means a data engineer here is wiring together product telemetry, fab yield signals, competitive pricing feeds, and increasingly, AI accelerator performance benchmarks across the full MI300 and Radeon lines. The stack confirms this: Snowflake, Databricks and Iceberg shows a warehouse-first architecture leaning on Snowflake for analytical serving, Databricks for heavier compute, and Iceberg signaling a push toward an open table format that can survive schema churn across long hardware cycles. SQL, Python and JavaScript rounds out to SQL, Python, and JavaScript, suggesting pipelines feed both internal analysts and some product-facing surfaces. The job leans toward batch and modeling work over low-latency streaming, though competitive intelligence and supply chain visibility can create real SLA pressure on freshness.