Compiler Verification Engineer, Compute Performance – GPU

NVIDIA•Austin, TX

About The Position

NVIDIA is seeking a Verification Engineer to join its Compiler Verification Team, focusing on compute performance and GPU compilers. The role analyzes compiler performance degradation and functional defects, identifies regression root causes, recommends corrective actions, and reviews testing to support continuous improvement. The engineer automates compiler testing with NVIDIA frameworks and programming, including test execution, reporting, results analysis, and automation of build and test environments. They work with compiler developers to provide automated unit-testing solutions and use test suites to detect, report, and track performance changes. The role also investigates performance data and outliers, monitors trends, and maintains historical data and baselines. Additional work includes developing and reviewing test plans, implementing and automating test cases, integrating tests into NVIDIA test-management frameworks, porting third-party tests, and authoring reports. The engineer follows iterative planning and test-development processes, identifies process weaknesses, proposes improvements, and participates in quality initiatives.

Requirements

Bachelor’s or Master’s degree, or equivalent experience.
At least 2 years of work experience in a software development or test organization.
Excellent communication skills; self-motivated and well organized.
Deep understanding of the Software Development Life Cycle (SDLC), High-Performance Computing (HPC), and software testing methodologies.
Understanding of how compilers work and are implemented, with proven problem-solving ability and ability to implement solutions.
Ability to collaborate with various teams to address performance regressions, work under tight schedules, and apply strong analytical skills and attention to detail.
Experience writing test plans and performing test development, test automation, test execution, and reporting in a production environment.
Programming and/or testing experience in C, C++, or CUDA, as well as scripting languages such as Python, Perl, or Shell.

Nice To Haves

Proven experience testing production software, preferably compiler or other system software.
Previous compiler development, compiler verification or testing, or performance analysis experience.
Experience with the NVIDIA CUDA Toolkit, especially solving issues and debugging in a Linux environment.
Background with revision control software and management tools such as Git, Perforce, JIRA, Confluence, and Make.
Familiarity with statistical analysis tools for identifying and isolating out-of-bound behavior.

Responsibilities

Analyze compiler performance degradation and functional defects, identify regression root causes, suggest corrective actions, and review testing.
Automate compiler testing, including test execution, reporting, results analysis, and build and test environment automation.
Work with compiler developers to provide automated unit-testing solutions.
Apply test suites to identify, report, and track compiler performance changes, and work with development teams to resolve regressions.
Generate statistics from performance data, investigate outliers, monitor performance trends, and maintain historical data and baselines.
Develop and review test plans; implement, automate, and integrate test cases into NVIDIA test-management frameworks.
Port third-party tests and author test reports, including integrating existing tests into compiler test automation.
Follow iterative planning and test-development processes, identify process weaknesses, propose improvements, and participate in quality initiatives.