Data Engineering at Goldman Sachs

Goldman Sachs runs some of the most latency-sensitive and correctness-critical data pipelines in any industry. Trading desk feeds, risk aggregations, regulatory reporting under Basel and Dodd-Frank, real-time position reconciliation: these are the data problems a DE here owns. The firm processes enormous transaction volumes across equities, fixed income, and derivatives, which means the engineering work centers on auditability and determinism as much as throughput. A pipeline that drops a row silently is not a performance issue at Goldman; it's a compliance event. The visible stack reflects that priority: tooling choices at the firm tend toward proven infrastructure over frontier technology, and the DE role tilts toward ownership of correctness guarantees, lineage tracking, and SLA adherence across systems that feed both internal risk models and external regulatory submissions.

Goldman's data reporting skews heavily toward mid-level: 80 of the 84 verified reports sit at L4, with 4 at entry. The firm is consistently filling its middle band, and candidates arriving with 5 to 8 years of experience in financial services data infrastructure are the clearest fit. Engineers who thrive here tend to be rigorous about correctness, comfortable in compliance-adjacent environments, and patient with slower decision cycles. If you're coming from a scrappier startup background and want to ship fast and own the roadmap, the cultural fit is likely poor regardless of technical ability. Engineers who should pass: anyone whose primary motivation is equity upside or tooling novelty. For anyone else, prep the Python data manipulation and SQL window function problems that the interview surfaces most, and make sure you can speak concretely to production incidents you've owned before you enter the loop.