About the job
As a Senior/Staff Engineer on the Foundation Model Compute Infrastructure team, you will lead the design and development of scheduling and orchestration systems for large-scale TPU workloads across multi-region clusters. You will work on distributed systems that manage thousands of accelerators and enable reliable, efficient execution of large-scale training and inference jobs. This role spans scheduling algorithms, cluster lifecycle management, workload orchestration, reliability engineering, and performance optimization.
Responsibilities
Design and evolve large-scale scheduling systems for TPU-based training and inference workloads across multi-region clusters
Build topology-aware, quota-aware, and fault-tolerant schedulers to improve utilization, fairness, startup latency, and reliability
Develop orchestration systems for distributed ML workloads running on Kubernetes and accelerator infrastructure
Improve cluster efficiency and operational scalability through automation of provisioning, resource management, quota workflows, and recovery handling
Collaborate closely with foundation model teams to support advanced distributed training and inference frameworks such as Pathways, Ray, and JAX-based workloads
Mentor engineers and influence architectural direction across Apple’s distributed AI compute platform
Qualifications
Minimum
7+ years of industry experience building large-scale distributed systems or cloud infrastructure
Strong programming skills in Python, Go, C++, or similar systems languages
Extensive experience with compute infrastructure and workload scheduling
Strong expertise in distributed systems, scalability, reliability, and performance engineering
Experience with Kubernetes, container orchestration, or large-scale cluster management systems
Experience designing backend services or infrastructure platforms operating at production scale
Strong communication and collaboration skills across engineering and research teams
Bachelor’s degree in Computer Science, Engineering, or related field
Preferred
Experience building schedulers, resource managers, or orchestration systems for distributed workloads
Experience with accelerator infrastructure such as TPU, GPU
Experience with distributed ML training or inference systems
Familiarity with frameworks such as JAX, PyTorch, TensorFlow, Ray, Pathways
Experience operating large-scale multi-tenant infrastructure in cloud or hybrid environments
Background in performance optimization, fault tolerance, or resource efficiency for large distributed systems
MS or PhD in Computer Science, Engineering, or related field