Job description

NVIDIA leads groundbreaking work in AI, high-performance computing, and visualization — and the memory subsystem is one of the most consequential parts of every silicon program we ship. GDDR, LPDDR, and HBM are where architecture, silicon, packaging, and manufacturing collide with reality. When memory is marginal, the whole system is marginal, and when it works, our GPUs and SoCs can hit the performance our customers depend on! We are hiring a Senior Memory Qualification Engineer inside the Silicon Co-Design Group who will own high-speed memory bring-up, characterization, and qualification on our flagship GPU, CPU and SoC programs. This is not a compliance-testing role — it is the seat where memory issues get root-caused before they become customer escapes. The two highest-leverage problems in this seat: • Root-cause memory marginality before software hardens. Isolate ambiguous, cross-boundary failures across DRAM, controller, PHY, IO tuning, board, PVT, and workload — and close them with productized fixes and reusable playbooks so the next program starts from a better baseline. • Build memory qualification as a repeatable, AI-accelerated capability. Every program generates large datasets of shmoo, margin, and characterization data. Turn that data into decision-ready insight faster than any single engineer can, and shift memory-issue detection left so we catch problems before they hit ramp! What you will be doing • Bring up high-speed memory interfaces (GDDR, LPDDR, HBM) on complex GPU, CPU and SoC silicon — from first power-on through full PVT characterization and productization. • Own functional validation, IO tuning, timing margin analysis, and multi-vendor DRAM qualification across the memory controller, PHY, and DRAM device — for the full matrix of capacities, speeds, and configurations we ship. • Lead root-cause closure on the hardest multi-team memory failures — spanning silicon, controller, PHY, package, board SI/PI, firmware, and workload — with reusable workarounds and productized fixes. • Design and deploy AI-enabled workflows for margin data analysis, shmoo bucketing, failure signature detection, and workload-vs-margin correlation — with the guardrails to make the outputs trustworthy enough to gate qualification decisions. • Partner with pre-Si architecture, design, and DV teams to translate new memory features into targeted post-Si validation plans, and feed post-Si learnings back into pre-Si methodology. • Work across globally distributed teams — silicon design, PHY, PCB, firmware, software, DRAM vendors, and manufacturing partners — including counterpart engineers in Taiwan and other major hubs. What we need to see: • B.Tech or M.Tech in Electrical, Electronics, or Computer Engineering (or equivalent experience), plus 10+ years in the semiconductor industry with at least 3 years directly on memory interface bring-up, characterization, and qualification on shipped silicon. • Deep protocol and physical-layer understanding of GDDR, LPDDR, or HBM — training, calibration, IO tuning, timing margin, and vendor-to-vendor behavior differences — with at least one specific example of taking an ambiguous memory failure to root-cause closure with a productized fix. • Strong hands-on lab fundamentals — oscilloscope, DMM, thermal chambers, protocol analyzers — combined with working knowledge of PCB stack-up, board layout, power planes, and SI/PI guidelines that show up as real memory issues. • Demonstrated AI-driven workflow you built or scaled for memory characterization, margin analysis, or failure triage — with adoption beyond yourself and measurable impact on debug velocity, coverage, or escape rate. Ways to stand out from the crowd: • Experience resolving ambiguous memory failure from symptom to root cause without being handed the answer; codified the fix as methodology adopted beyond their immediate program. • A history of building reusable qualification methodology, debug playbooks, or margin-analysis frameworks that were adopted by other programs or teams . • AI work that goes beyond personal-copilot use — agentic workflows, RAG-grounded debug assistants, automated shmoo analysis, ML-driven margin prediction — deployed at team scope with adoption metrics. NVIDIA is the world leader in accelerated computing, and our work powers AI, gaming, robotics, autonomous systems, and scientific discovery. We invest in our people with competitive benefits, flexible time off, and continuous learning, and we build a team where everyone can do their best work. Your base salary will be determined based on your location, experience, and the pay of employees in similar positions. The base salary range is 168,000 USD - 264,500 USD for Level 4, and 196,000 USD - 310,500 USD for Level 5. You will also be eligible for equity and benefits . Applications for this job will be accepted at least until October 5, 2026. This posting is for an existing vacancy. NVIDIA uses AI tools in its recruiting processes. NVIDIA is committed to fostering an inclusive work environment and proud to be an equal opportunity employer. As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.

Requirements

Must-haves

What you will be doing
Bring up high-speed memory interfaces (GDDR, LPDDR, HBM) on complex GPU, CPU and SoC silicon — from first power-on through full PVT characterization and productization.
Lead root-cause closure on the hardest multi-team memory failures — spanning silicon, controller, PHY, package, board SI/PI, firmware, and workload — with reusable workarounds and productized fixes.
Partner with pre-Si architecture, design, and DV teams to translate new memory features into targeted post-Si validation plans, and feed post-Si learnings back into pre-Si methodology.
Work across globally distributed teams — silicon design, PHY, PCB, firmware, software, DRAM vendors, and manufacturing partners — including counterpart engineers in Taiwan and other major hubs.
Demonstrated AI-driven workflow you built or scaled for memory characterization, margin analysis, or failure triage — with adoption beyond yourself and measurable impact on debug velocity, coverage, or escape rate.
Experience resolving ambiguous memory failure from symptom to root cause without being handed the answer; codified the fix as methodology adopted beyond their immediate program.
A history of building reusable qualification methodology, debug playbooks, or margin-analysis frameworks that were adopted by other programs or teams .
AI work that goes beyond personal-copilot use — agentic workflows, RAG-grounded debug assistants, automated shmoo analysis, ML-driven margin prediction — deployed at team scope with adoption metrics.
You will also be eligible for equity and benefits .
Applications for this job will be accepted at least until October 5, 2026.
This posting is for an existing vacancy.
NVIDIA uses AI tools in its recruiting processes.

Nice-to-haves

None flagged.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aiml

Platform: eightfold
Effort: medium

Benefits

Benefits detected in the listing.

Pto