Senior Machine Learning Engineer - Orchestration

Responsibilities

• Optimize distributed scheduling and orchestration strategies to improve cluster utilization and resource efficiency.
• Develop and integrate autoscaling, automatic parallelization, resource preemption and eviction, mixed-resource deployment, and scheduling across data centers, regions, and clouds.
• Build distributed training runtimes for ultra-large-scale embeddings and GPU synchronization.
• Design distributed computing APIs and runtime systems for recommendation and advertising model workflows including reinforcement learning, fine-tuning, and distillation.
• Improve the diagnosability and usability of distributed training platforms.
• Build distributed online model inference architecture for large-scale embedding and recommendation workloads.
• Improve recommender-system online architecture and ML operations processes by integrating business research and experimentation models.

Requirements

• Bachelor's degree or above in Computer Science or a related field.
• At least 5 years of experience and strong hands-on proficiency in Go or Python in a Linux environment.
• Experience with distributed scheduling frameworks such as Kubernetes, Godel, YARN, Mesos, or Celery, and with machine learning systems.
• Experience designing, developing, and maintaining large-scale distributed systems, with a strong understanding of distributed-systems principles.
• Strong logical analysis, business abstraction, communication, learning, documentation, and execution skills.
• Preferred experience with PyTorch or TensorFlow.
• Preferred experience in AI infrastructure, hardware/software co-design, high-performance computing, or ML hardware architecture involving GPUs, accelerators, or networking.
• Preferred experience designing or using training orchestration systems such as veRL, vLLM, Ray, or TFX.