Software Engineer, Data Systems

NVIDIA•Santa Clara, CA

About The Position

NVIDIA is seeking a Software Engineer, Data Systems to join its team in Santa Clara, California. The role focuses on building software that helps technical users understand complex computing systems and analyze AI workload performance across GPU data centers. The engineer will manage the data store, data model, and query path, turning large streams of profiling and telemetry data into useful information for engineers and automated analysis. The position involves designing and operating data systems that ingest, join, store, and deliver engineering performance data at GPU data-center scale with low latency. It also includes developing data APIs and evidence views, managing data contracts with producing teams, and balancing data quality, query latency, storage cost, and retention. The role requires making data limitations visible to downstream users and maintaining secure, stable service behavior under high demand or failure. The base salary range is $224,000–$356,500. The employee will also be eligible for equity and benefits. Applications will be accepted at least until November 2, 2026. This posting is for an existing vacancy.

Requirements

Bachelor’s degree in Computer Science, Computer Engineering, or a related field, or equivalent experience.
12+ years of relevant software engineering experience, including delivery and operation of data pipelines, storage systems, analytical services, or comparable software used by others.
Experience handling large-scale production data, including billions of events per day or petabyte-class stores.
Experience with low-latency streaming pipelines that manage backpressure, replay, and duplicate delivery accurately.
Good judgment about event identity and ordering, data transformations, and how errors or missing data affect results.
Ability to explain a concrete performance investigation, including its baseline, measurements, competing explanations, changes made, verified results, and trade-offs.
Experience establishing and versioning a data contract with a producing team and guiding consumers through a format change.
Ability to debug across component boundaries and validate changes through code review, tests, and independent measurements, including when using automated development tools.
Clear communication about design choices, observed behavior, and uncertainty, with care for users’ data and operational constraints.
Proficiency with C++, Python, and SQL.

Nice To Haves

Hands-on experience with telemetry, time-series data, or analytical query engines such as ClickHouse, and columnar formats such as Arrow and Parquet.
Familiarity with profiler data, including trace record formats, sampling semantics, symbolization, clock domains, and aligning events recorded on different machines.
Experience building high-cardinality schemas and measuring improvements to memory use, storage efficiency, or query performance while preserving correctness.
A track record of making data processing easier to inspect and recover when inputs change, traffic spikes, or failures occur.
Working knowledge of Go.

Responsibilities

Develop and construct the store, data model, and query path for engineering performance data, including ingesting, joining, storing, and delivering it at GPU data-center scale with low latency.
Join profiling data with system context such as hosts, schedulers, configuration, logs, and storage, ensuring records have the identity and dimensions needed for comparisons across jobs, machines, versions, and time.
Establish and version data contracts with data-producing teams, and transition consumers through changes without disruption.
Balance data quality, query latency, storage cost, and retention; measure and verify improvements in efficiency, latency, and resource use.
Handle delayed, duplicated, incomplete, or inconsistent data and make its limitations visible to downstream users.
Develop data APIs and generated evidence views for people and automated analysis.
Implement access controls and resource limits, and maintain stable behavior during high demand or failures.