Performance Engineer, Low-Level Libraries

Meta•Menlo Park, CA

8d•$85 - $251,000

About The Position

Are you committed to squeezing every drop of performance? Join Meta's Low-Level Libraries team and drive impact across our foundational infrastructure. We own the performance and Developer Experience (DevX) for critical C/C++ libraries like folly, Jemalloc, and GEMM (MLK, Aocl-Blas, etc.). Our work involves cutting-edge optimizations in domains such as memory management, concurrency, architecture-specific enablement, and AI frameworks. The result? Significant power savings and enabling new platforms at Meta.

Requirements

Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
7+ years of professional C/C++ experience
Knowledge of computer architecture, CPU and memory subsystem, and OS-level resource management
Experience using performance-profiling tools and optimizing native applications for execution-time and memory efficiency

Nice To Haves

Experience implementing and optimizing low-level libraries, such as memory management, threading, GEMM, data compression, or string processing
Knowledge of modern ISAs, such as x86 and ARM
Experience hand-tuning code, e.g., with loop optimizations, vectorization, parallelization, HW-architecture-specific optimizations
Experience developing operating-system kernels

Responsibilities

Develop and optimize C/C++ libraries for Meta services–memory allocation, thread pools and work scheduling, thread synchronization and lockless data structures, highly performant collections, async processing and I/O, RPC, etc
Analyze resource utilization in server applications (CPU, GPU, memory, network, etc.), identify bottlenecks, scope out opportunities for improved resource utilization, and implement improvements, such as modifying core libraries to optimize Meta server workloads, implementing efficiency improvements in production code (e.g., change core data structures), or improving server utilization
Work with internal customers and partners to define requirements
Reflect requirements in the team roadmap and plan out execution