Staff -Power and Performance Validation Engineer

About the role

Graphcore develops hardware and software infrastructure to advance artificial intelligence compute and support global AI adoption. As a SoftBank Group company, the organization focuses on creating technology that enables artificial super intelligence. The role centers on ensuring that Graphcore silicon and platforms meet stringent power, thermal, and performance targets before deployment. You will own the end to end validation strategy that de risks the architecture and confirms system behavior under real world constraints. This position sits within the Architecture and Validation team where you will act as the bridge between design intent and silicon reality. You will apply rigorous testing methodologies to validate that hardware features perform as specified at scale and in production conditions. The work demands a methodical mindset that balances deep technical analysis with clear communication of results.

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 validation tasks to ensure correct hardware initialization and stable operation.
• Build automation frameworks, regression systems, and reporting tools using Python to increase test coverage and repeatability.
• Test CPU, memory, interconnect, and high-speed I/O subsystems under various workload conditions to expose integration issues.
• Design benchmark workloads and parameter sweeps that explore system configuration space and uncover optimization paths.
• Coordinate with RTL, firmware, software, and architecture teams to troubleshoot complex technical problems that span multiple domains.
• Establish validation metrics and pass fail criteria to ensure data quality and consistency across test campaigns.
• Develop custom micro-benchmarks and workload generators that target specific hardware features or bottlenecks.
• Present technical findings and project risks to engineering leadership to support informed decision making.
• Instrument test setups to capture telemetry, power traces, and performance counters with high precision.
• Maintain detailed test logs and results that support root cause analysis and long term product insights.
• Collaborate on the definition of test plans that align with product milestones and risk mitigation goals.
• Drive test execution on automated fixtures while monitoring for anomalies in real time.
• Contribute to the continuous improvement of validation infrastructure and methodology documentation.

Requirements

• Professional experience in silicon characterization, system performance engineering, or power and performance validation within a semiconductor or hardware environment.
• Knowledge of modern SoC architecture including compute, memory hierarchy, interconnect, and high-speed I/O subsystems and their interaction with software.
• Proficiency in Python for data analysis, test orchestration, automation scripting, and result visualization.
• Experience with Linux systems, shell scripting, and low level performance analysis using standard profiling tools.
• Familiarity with profiling and benchmarking tools such as iperf, perf, fio, and stress-ng to design meaningful experiments.
• Ability to debug intricate hardware and software interactions across multiple layers from firmware through applications.
• Capability to design structured validation methodologies, test strategies, and measurement frameworks that scale.
• Strong attention to detail and ownership of test quality, with the discipline to reproduce issues and track outcomes systematically.

Nice to have

• Background in high performance compute architectures, GPUs, or AI accelerators and their associated workload characteristics.
• Experience with pre-silicon simulation or modeling environments that help correlate expected and observed behavior.
• Understanding of silicon characterization methodologies and dynamic power management technologies used in modern processors.
• C or C++ programming skills for developing benchmarks, low level system utilities, or performance instrumentation.
• Familiarity with telemetry systems and hardware instrumentation including oscilloscopes, power sensors, and protocol analyzers.
• Experience with large scale or high core count compute systems and the challenges of scaling benchmarks across many devices.
• History of contributing to engineering best practices, test frameworks, and shared validation infrastructure that benefits the wider organization.

Practical notes

This is a full time position based in Bengaluru, India. The role may require travel within India for customer engagements, vendor visits, or technical workshops as project needs dictate. Candidates must be eligible to work in India without visa sponsorship for this role. There are no salary details provided in the source information. The successful candidate will join a growing team responsible for validating next generation AI silicon in a demanding production environment.