Senior Power & Performance Validation Engineer

About the role

Graphcore is a company that develops specialized hardware and software infrastructure to advance artificial intelligence compute. As a subsidiary of the SoftBank Group, the company focuses on creating systems that support the development of artificial super intelligence. In this role, you will own the end-to-end validation of power and performance for next generation AI silicon and platforms. You will create and manage comprehensive strategies for thermal, power, and performance characterization across the entire product stack. You will oversee post-silicon bring-up and characterization phases to ensure systems meet strict targets before deployment. You will build and maintain automated frameworks and regression systems using Python to drive efficiency and repeatability. You will evaluate critical subsystems including memory, interconnects, high-speed I/O, and CPUs under varied and demanding workloads. You will design micro-benchmarks and custom workload generators to test system limits and uncover hidden behaviors. You will analyze performance data to pinpoint bottlenecks and suggest architectural or firmware improvements for optimization. You will coordinate closely with RTL, firmware, and software teams to troubleshoot complex hardware-software integration issues. You will establish validation metrics and reporting standards to ensure product reliability and consistency across deployments.

Key facts

What you'll do

• Create and manage strategies for thermal, power, and performance characterization of AI silicon and platforms.
• Oversee post-silicon bring-up and characterization phases to validate functionality under real conditions.
• Build automated frameworks and regression systems using Python to streamline testing and data collection.
• Evaluate subsystems including memory, interconnects, high-speed I/O, and CPUs under varied workloads to assess stability.
• Design micro-benchmarks and custom workload generators to test system limits and drive innovation.
• Analyze performance data to pinpoint bottlenecks and suggest architectural or firmware improvements for gains.
• Coordinate with RTL, firmware, and software teams to troubleshoot complex hardware-software integration issues.
• Establish validation metrics and reporting standards to ensure product reliability and consistency across teams.
• Drive test planning and execution for power and performance validation across multiple platforms and configurations.
• Collaborate with cross-functional teams to define characterization goals and align with product roadmaps.
• Implement scalable test methodologies that improve coverage while reducing time to insight.
• Leverage system-level knowledge to validate interactions between compute, memory, and I/O subsystems.
• Support the development of tools and scripts that automate analysis and visualization of large data sets.
• Contribute to the creation of technical documentation that captures methodologies and results for future use.

Requirements

• Professional background in silicon characterization, system performance engineering, or power and performance validation.
• Technical knowledge of modern SoC architectures, specifically memory, CPUs, interconnects, and high-speed I/O.
• Proficiency in Linux environments and low-level performance analysis to investigate issues in depth.
• Advanced Python skills for data analysis, orchestration, and automation of test workflows.
• Practical experience with profiling and benchmarking tools such as iperf, perf, fio, or stress-ng.
• Ability to design structured test strategies and interpret large technical datasets to derive conclusions.
• Experience resolving complex hardware and software interaction bugs to ensure system stability.
• Strong understanding of performance metrics, power states, and thermal behavior in compute systems.
• Demonstrated capability to work independently and manage multiple priorities in a fast paced environment.
• Willingness to follow strict procedures while also innovating within validation frameworks.
• Excellent written and verbal communication skills for clear reporting and collaboration.
• Attention to detail to ensure accuracy in measurement, reporting, and data integrity.
• Commitment to learning new technologies and adapting to evolving hardware generations.

Nice to have

• Background in high-performance compute, GPU, or AI accelerator architectures.
• Experience with pre-silicon simulation or modeling environments to guide validation planning.
• Familiarity with telemetry systems and hardware instrumentation for data collection.
• Knowledge of silicon characterization and power management techniques to optimize behavior.
• C or C++ programming skills for benchmark development or low-level systems work.
• Experience working with large-scale or high core-count compute systems in production settings.

Practical notes

• Location is restricted to Bengaluru, India.
• Engagement is full-time.
• Travel may be required depending on project needs and team coordination.
• Visa sponsorship may be considered for eligible candidates based on role requirements.
• Candidates must be able to start within the timeframe specified during selection.
• This role operates within the policies and guidelines defined by Graphcore and its subsidiaries.