Power and Performance Validation Engineer, Annapurna Labs
Core
Own power-domain validation for next-generation machine learning accelerators across the product lifecycle, from silicon bring-up to production support in AWS data centers.
Role type
Power and Performance Validation Engineer (Silicon)
Builds
Custom silicon ML accelerators for AWS cloud infrastructure
Domain
Semiconductor hardware engineering / Machine Learning hardware
Deliverable
production ML models
Required skills
Power characterization, power delivery validation, DVFS flows, peak power management, di/dt handling, power measurement instrumentation, data analysis for power optimization
Preferred skills
Power integrity simulation (ANSYS RedHawk, Cadence Voltus), ML hardware/software architecture, CI/CD, voltage regulator design, power management firmware validation
Technologies
DC power analyzers, current probes, high-bandwidth oscilloscopes, VNA/TDR, C++, Python, Golang
Responsibilities
Develop power validation strategies and test plans covering sequencing, DVFS, peak power, and stress testing; Characterize peak power consumption and transient droop under real ML workloads; Conduct hands-on power measurements and debug using lab instrumentation; Validate power management algorithms and thermal throttling mechanisms; Correlate silicon measurements with pre-silicon models; Build automated power regression frameworks and dashboards; Collaborate with architecture and firmware teams to triage power anomalies; Support production systems and address field power issues
Seniority
Mid-level, hands-on IC
