Software Engineer, Accelerator Solutions & Technologies (PhD)

Job Title

Software Engineer, Accelerator Solutions & Technologies (PhD)

Role Summary

Engineer on the Accelerator Solutions & Technologies team focused on collective communications software, simulator/emulation infrastructure, and performance optimization for Meta’s accelerator-based AI/ML superclusters.

The role works with cross-functional hardware, firmware, and architecture teams to model, simulate, benchmark, and improve distributed training and inference performance at datacenter scale.

Experience Level

Mid-level — expects several years of practical software development experience (posting lists 2+ years of development experience in key languages and related systems work).

Responsibilities

Core responsibilities include building and validating simulation/emulation platforms, analyzing performance, and integrating results into designs.

• Develop and maintain simulation and hardware emulation platforms for accelerator performance measurement and optimization.
• Contribute to collective communications library design and implementation for large-scale AI/ML clusters.
• Support networking and hardware acceleration techniques to improve distributed training and inference throughput and latency.
• Perform architectural analysis to validate system performance, scalability, and reliability targets.
• Implement simulation models for accelerators, run scenario analyses, and identify system-level improvements.
• Use instruction set simulators to specify and validate accelerator firmware behavior.
• Collaborate with hardware and firmware teams to ensure simulation fidelity and integrate findings into product designs.
• Analyze simulation and benchmark data to guide firmware and software optimizations.

Requirements

Must-have technical skills and experience; preferred items listed separately.

• 2+ years developing and maintaining C++ codebases.
• 2+ years developing Python codebases.
• Experience with performance measurement, benchmarking, and optimization for distributed model training and collective communications.
• Experience building or using simulation/emulation infrastructure and interpreting simulation results for system design.
• Ability to collaborate with architects, hardware, and firmware engineers to translate simulation results into design changes.

Nice-to-have

• Familiarity with transport stacks and interconnect constraints (e.g., RoCE).
• Experience with SystemC or similar hardware modeling frameworks.
• Knowledge of AI/HPC hardware requirements and experience with frameworks/tools such as PyTorch and CUDA.
• Experience applying responsible/ethical AI practices and integrating AI tools to optimize workflows.

Education Requirements

Posting requires a PhD in Computer Science, Computer Engineering, or a relevant technical field (or equivalent practical experience) and also references a Bachelor's degree in those fields; degrees must be completed prior to joining Meta. Equivalent practical experience is accepted per the posting.

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.