GPU Performance Modelling Architect

Drive architecture exploration and performance analysis for next-generation GPUs through modeling, simulation, and workload characterization. Partner with architecture, software, compiler, and hardware teams to influence design decisions using data-driven insights.

Key Deliverables

- Develop and maintain GPU performance models for architecture evaluation.

- Conduct Performance analysis of graphics and compute workloads to identify bottlenecks.

- Perform architectural trade-off studies to generate and defend performance projections.

- Bring up new workloads on Performance Models by fixing bugs and developing new features.

- Evaluate new features and provide recommendations to improve performance and efficiency.

- Build automation and analysis tools to improve modeling productivity.

Responsibilities

- Design and execute performance experiments and simulations.

- Conduct root-cause analysis and derive actionable architectural insights.

- Collaborate with cross-functional teams on feature development and optimization.

- Develop tools, workflows, and methodologies for performance analysis.

- Present findings and influence future architecture decisions.

Critical / Must-Have Skills

- Strong understanding of computer architecture

- Strong analytical, debugging, and problem-solving skills.

- Proficiency in C/C++ and Python.

- Experience with architectural simulators and performance prediction.

- Experience with performance analysis, profiling, and workload characterization.

- Strong verbal and written communication skills

Good-to-Have Skills

- GPU/CPU/Memory architecture and/or Graphics fundamentals.