Research Scientist, ML H/W-SW Codesign

Job Title

Research Scientist, ML H/W-SW Codesign

Role Summary

Join Reality Labs Silicon to develop ML model and hardware co-design solutions that enable compute- and power-efficient training and on-device inference for AR/VR and edge devices. The role involves cross-disciplinary research spanning algorithms, hardware architecture, firmware, and software integration within an R&D research team.

Experience Level

Entry-level / Early career. Requires at least 2 years of specialized experience in relevant ML/hardware domains.

Responsibilities

Primary responsibilities include cross-discipline research, hardware/software co-design, model optimization, and benchmarking.

• Identify and solve multi-discipline ML acceleration problems across algorithms, network design, hardware architecture, and AR/VR use cases.
• Collaborate across hardware and software teams to address deep co-design problems.
• Codesign and invent ML accelerator and system architecture solutions and integrate algorithm and software support.
• Develop model compression and scalability techniques (numerics, pruning, distillation, etc.).
• Optimize models for hardware accelerators to meet real-time latency and power constraints.
• Define use cases and develop benchmarks and methodologies to evaluate approaches.
• Perform thorough, data-driven analysis to influence product and engineering decisions.
• Stay current with research, attend conferences, and publish or patent novel outcomes.

Requirements

Must-have qualifications and experience are listed first; preferred items are labeled "Nice-to-have."

• 2+ years of specialized experience in one or more of: model compression, hardware-aware model optimizations, accelerator architecture, GPU architecture, ML compilers, ML systems, AI infrastructure, high-performance computing, performance optimization, or ML frameworks (e.g., PyTorch).
• Experience developing AI system infrastructure, AI algorithms, or AI hardware acceleration in C/C++ or Python.
• Authorization to work in the country of employment at the time of hire and maintain ongoing work authorization during employment.

Nice-to-have:

• Experience with training or inference of large-scale AI models (computer vision and/or LLMs).
• Experience architecting ML hardware accelerators and systems.
• Experience with PyTorch, TensorFlow, or similar toolsets.
• On-device algorithm development experience and optimizing ML compilers for efficient accelerator deployment.
• Track record of publications, patents, grants, or notable open-source contributions.
• Experience working and communicating cross-functionally in a team environment.

Education Requirements

Master's or PhD in Electrical Engineering, Computer Science, or equivalent practical experience. Related technical fields and equivalent experience are accepted.

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.