3 hours ago
Base Salary
$217k - $303k/yr
Responsibilities
- Own production ML performance optimization initiatives from bottleneck identification through implementation, validation, and rollout.
- Profile and diagnose bottlenecks across model execution, frameworks, runtimes, memory, CPU/GPU utilization, and infrastructure.
- Implement improvements involving inference execution, runtime behavior, batching, memory management, precision, compiler integration, custom operators, and kernels.
- Optimize ML training and serving for latency, throughput, hardware utilization, cost, model quality, and reliability.
- Support model launches through benchmarking, load testing, regression detection, and launch-readiness improvements.
- Partner with model engineering, serving, ranking, and ML Platform teams to integrate and validate optimizations.
- Build reusable performance benchmarks, tooling, optimization patterns, and engineering guidance.
- Contribute to technical direction and mentor engineers through collaboration, code reviews, debugging, and knowledge sharing.
Requirements
- Demonstrated hands-on experience diagnosing and implementing meaningful performance improvements in production or production-relevant ML workloads.
- Deep expertise in at least one area such as inference execution, ML runtimes, compiler integration, GPU/kernel optimization, memory systems, or training execution.
- Strong profiling and benchmarking skills, including establishing baselines and validating measurable end-to-end improvements.
- Strong software engineering fundamentals, including implementing, debugging, testing, and maintaining performance-sensitive systems with proficiency in Python and relevant systems programming technologies.
- Understanding of production ML frameworks and trade-offs across model execution, runtime, memory, hardware, and infrastructure.
- Ability to balance performance with numerical correctness, model quality, reliability, maintainability, and safe rollout.
- Clear communication and ability to explain technical trade-offs to engineers and partner teams.
- Experience optimizing large-scale inference or training workloads in recommendation systems, ranking, search, or retrieval is preferred.
- Experience with PyTorch performance optimization, distributed-training performance, inference runtimes, GPU execution, hardware-aware optimization, benchmarks, profiling tools, regression detection, or ML cost observability is preferred.
- Specialized compiler internals or custom CUDA/Triton kernel development are valuable but not required.
Benefits
- Remote work is available from anywhere in the United States or Canada under Reddit’s flexible first workforce model.
- Global benefits include workspace support, professional development, and caregiving support.
- Family planning support, gender-affirming care, mental health and coaching benefits, medical and dental coverage, income replacement programs, and paid parental leave are offered.
- Additional benefits include a group personal pension scheme with employer match, bike-to-work scheme, flexible vacation, and paid volunteer time off.
- U.S.-based employees may receive medical, dental, and vision insurance, a 401(k) program with employer match, and generous vacation and parental leave.
- The role may include equity in the form of restricted stock units and, depending on the position offered, commission.
Categories
About Reddit
Reddit builds a social news and discussion platform organized into user-created forums (“subreddits”) for consumers, creators, and communities. It monetizes primarily through advertising tools for brands and self-serve advertisers, plus a Premium subscription; developers can access data via its API. Founded in 2005 and headquartered in San Francisco, Reddit became a public company on the NYSE in 2024 and hosts over 100,000 active communities.