Research Scientist, HW/SW Co-Design (PhD) at Meta, Menlo Park, CA

Mission:

Explore, develop, and help productionize high performance software & hardware technologies for AI at datacenter scale.
Concurrent design and optimization from models and runtime to AI hardware.
Optimize across compute, network, and storage.
Invest in model optimization on existing accelerator systems and guide future models and AI hardware at Meta.
Impact key AI services: Generative AI and Recommendations.
Collaborate with industry, academia, Meta’s Infrastructure and Product groups.
From ideation and rapid prototyping to productization of high leverage ideas.

Achievements:

Team members have won Best Paper Awards at ISCA, ASPLOS, SOSP, OSDI.
Multiple papers selected for IEEE Micro Top Picks.
Publications in ICML, NeurIPS, SC, HPCA, NSDI, VLDB, MLSys.
Research communities: systems for ML (MLSys, SOSP, OSDI), hardware architecture (ISCA), ML (NeurIPS), supercomputing (SC).

Role:

Focus on research & development at intersection of Generative AI workload analysis, model enablement, co-designing architecture for Meta's custom AI accelerators.
Explore theoretical underpinnings and practical implementation of novel hardware-aware mapping techniques.

Responsibilities:

Pioneer hardware-software co-design for Meta's custom AI silicon focusing on programmability, performance, power efficiency.
Integrate new silicon/system technologies into custom AI accelerator roadmap based on workload analysis and future model/GenAI requirements.
Build system performance models/simulators for datacenter infrastructure options analysis.
Co-optimize deep learning kernels/primitives with hardware architects/internal compiler teams for maximum efficiency on Meta's hardware platform.
Influence hardware roadmap of custom AI accelerators.
Architect/implement advanced frameworks/tooling for comparative analyses across system architectures.
Lead cross-functional initiatives spanning multiple engineering organizations to drive high-impact milestones.
Publish research results in recognized conferences (NeurIPS, ICML, ICLR, ASPLOS, ISCA, HPCA, MLSys).

Minimum Qualifications:

PhD in Computer Science/Electrical Engineering/Applied Mathematics or relevant field or equivalent practical experience (completed prior to joining).
Theoretical background/practical experience with AI models (CNNs, Transformers, LLMs, Diffusion Models).
Research experience in hardware-aware model enablement/performance modeling of AI systems or accelerators/silicon architectures.
Experience in system-level performance analysis/profiling/benchmarking of AI workloads.
Hands-on proficiency with end-to-end AI hardware architecture or on-device mapping algorithm development including logic/architecture/optimizations for performance/power/area (PPA).
In-depth Python experience and experience with at least one major AI framework.
Track record of publishing research in peer-reviewed venues; ability to communicate technical results to technical/nontechnical stakeholders.
Work authorization required.

Preferred Qualifications:

Experience with state-of-the-art performance modeling infrastructure and contributing to AI model performance projections.
Proven track record demonstrated by grants/fellowships/patents/first-authored publications at leading workshops/conferences relevant to position.
Experience deploying AI agents and techniques for increased efficiency.
Familiarity with low-level programming for specialized hardware (CUDA/HIP/Triton) or hardware description languages (HDL).
Experience co-designing hardware/software interfaces or influencing custom silicon architecture.
Experience working/communicating cross-functionally in team environment.

About Meta: Meta builds technologies that help people connect/find communities/grow businesses. Moving beyond 2D screens toward immersive experiences like AR/VR. Careers at Meta shape a future beyond digital connection constraints. Compensation: $122K - $181K/year + bonus + equity + benefits. Individual compensation based on skills/qualifications/location. Base salary excludes bonus/equity/incentives. Equal Employment Opportunity: Meta is an Equal Employment Opportunity employer; does not discriminate based on legally protected characteristics. Provides reasonable accommodations for qualified individuals with disabilities. Application: Apply now link provided.