Research Scientist, Systems ML - HW/SW Co-Design

Job Title

Research Scientist, Systems ML - HW/SW Co-Design

Role Summary

Member of the AI & Systems Co-Design (MTIA Co-Design) team responsible for translating production ML workload behavior into accelerator architecture decisions. Work spans workload characterization, performance modeling, micro-architecture definition, and pre-silicon validation.

Collaborate with silicon design, ML infrastructure, and product teams to influence hardware roadmap choices that improve performance, power efficiency, and cost for large language models, recommendation systems, and generative AI at hyperscale.

Experience Level

Senior level - requires significant industry experience; minimum 7+ years of relevant industry experience (or equivalent).

Responsibilities

Contribute to architecture, evaluation, and cross-functional initiatives that guide pre-silicon decisions for next-generation AI accelerators.

• Translate production ML workload insights into hardware design recommendations to improve performance, power, and cost.
• Lead data-driven analysis and performance modeling of hardware micro-architectures to inform silicon investment decisions.
• Design and build tooling, benchmarks, and evaluation frameworks for comparative architecture studies.
• Drive cross-functional programs spanning silicon design, ML infrastructure, and product teams.
• Define use cases, benchmarks, and evaluation criteria for hardware architecture assessment.
• Bridge ML systems knowledge with hardware, networking, and memory system design to enable novel architecture choices.
• Mentor researchers and engineers and establish standards for reproducibility, benchmarking, and technical rigor.

Requirements

Technical skills and experience required and preferred for success in this role.

Must-have

• 7+ years of industry experience in hardware/software co-design, AI accelerator architecture, systems for ML, high-performance computing, or performance modeling.
• Experience with power, performance, and area (PPA) trade-offs in hardware micro-architecture design.
• Strong understanding of modern ML workloads (large language models, generative AI) and how hardware choices affect their performance at scale.
• Contributed to at least one silicon tapeout, from architectural exploration through pre-silicon validation.
• Experience with AI system design considerations including networking, host-to-device ratios, and power trade-offs.
• Demonstrated ability to build performance models, benchmarks, and evaluation infrastructure for architecture decision-making.

Nice-to-have

• PhD in a relevant field.
• Experience with ML frameworks (e.g., PyTorch) and the full software stack from training/inference to hardware execution.
• Published research in top architecture or systems venues or equivalent industry contributions.
• Experience with numerics optimization (quantization, mixed-precision, custom number formats) and on-device algorithm/logic optimization.
• Track record of technical leadership, roadmap definition, and cross-team mentorship.
• Familiarity with responsible and ethical AI practices and integrating AI tools to optimize workflows.

Education Requirements

Bachelor's degree in Computer Science, Computer Engineering, Electrical Engineering, or a related technical field is listed as expected, or equivalent practical experience. A PhD in a related field is noted as preferred.

About the Company

Company: Meta Platforms

Headquarters: Menlo Park, California, United States

American technology company that develops social networking products (Facebook, Instagram, WhatsApp) and invests in virtual/augmented reality hardware and software through Reality Labs, focusing on connectivity, advertising, and immersive computing experiences.