Research Scientist, ML H-W/S-W Codesign

Job Title

Research Scientist, ML H-W/S-W Codesign

Role Summary

Develop and productize model–hardware co-design techniques for compute- and power-efficient training and on-device inference of vision and language models for AR/VR and edge devices within Reality Labs Silicon.

Work across algorithms, hardware architecture, and software systems to design ML accelerators, optimize models for constrained devices, and define evaluation methodology.

Experience Level

Mid-level research scientist. Advanced research and engineering experience in ML and system/hardware co-design expected.

Responsibilities

Primary responsibilities include inventing and evaluating co-design solutions and integrating them into systems.

• Identify and solve cross-discipline ML acceleration problems spanning algorithms, network design, hardware architecture, and AR/VR use cases.
• Co-design ML accelerators and system architectures and integrate algorithm and software changes to leverage hardware features.
• Develop model compression and scalability techniques (numerics, pruning, distillation) and optimize models for accelerators.
• Optimize for latency, power, and real-time constraints on target hardware.
• Define use cases, benchmarks, and evaluation methodology to compare approaches.
• Analyze trade-offs and communicate recommendations to partners and cross-functional teams.
• Stay current with research, contribute to publications and patents as appropriate.

Requirements

Must-have qualifications and relevant skills. Preferred items listed separately.

Must-have:

• Experience developing AI-system infrastructure, AI algorithms, or AI hardware acceleration using C/C++ or Python.
• Demonstrated ability to work across hardware and software to address co-design challenges for ML acceleration.
• Experience optimizing models for hardware accelerators to meet performance, latency, and power targets.
• Experience defining benchmarks and analyzing system-level trade-offs (performance, power, latency).

Nice-to-have:

• Experience with PyTorch, TensorFlow, or similar ML toolsets.
• Knowledge of training/inference of large-scale models (CV and/or LLMs) and on-device algorithm development.
• Experience architecting ML hardware accelerators or optimizing ML compilers for accelerators.
• Track record of research outputs: publications, patents, grants, or widely used open-source contributions.
• Experience implementing responsible, ethical AI practices and cross-functional communication.

Education Requirements

Listings reference a Bachelor's degree in Computer Science, Computer Engineering, or a relevant technical field and/or a PhD in Electrical Engineering or Computer Science. The posting also allows "equivalent practical experience."

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.