AIM Global Talent Hiring! Full Time AI Infra Platform Engineer in - Ricebowl

About the Opportunity

Our client is a fast-growing AI research and technology company building reasoning-first, agentic AI systems, with a footprint spanning the US and Asia. The team is behind several widely adopted open-source research agents that have posted top-tier results on industry benchmarks, and is led by scientific leadership with backgrounds spanning top US universities and frontier AI labs. Backed by a serial entrepreneur with a track record of building category-defining tech companies, the company is now scaling its compute infrastructure to support next-generation training and inference workloads at massive scale.

The Role

Build and evolve the core infrastructure layer for large-scale AI training and inference on 10,000+ GPU clusters — Kubernetes scheduling, storage, networking, and reliability engineering that makes massive shared compute efficient, reliable, and easy to operate for research and engineering teams.

What You'll Do

• Build and evolve Kubernetes-based GPU cluster infrastructure for large-scale AI training & inference
• Design and operate multi-tenant resource management: queue isolation, priority, quotas, preemption, batch orchestration, elastic resource allocation
• Improve deployment efficiency, workload stability, and overall GPU utilization across large shared clusters
• Manage high-performance storage for training data, checkpoints, and model artifacts — lifecycle management, access control, cost optimization
• Analyze & optimize networking and communication paths: RDMA, NCCL, bandwidth bottlenecks, cluster topology, cross-node communication efficiency
• Build cluster-level observability: logging, monitoring, alerting, and diagnostics
• Drive automation for cluster delivery, rollout, configuration management, fault handling, and day-2 operations
• Partner with training, inference, model, and platform teams to continuously improve the workload experience

What We're Looking For

• 5+ years in cloud-native infrastructure, distributed systems, ML platform engineering, or AI infrastructure
• Strong hands-on Kubernetes experience: container orchestration, cluster operations, scheduling, GPU resource management
• Solid understanding of large-scale GPU infrastructure challenges: scheduling, deployment, networking, storage, observability, reliability
• Familiarity with distributed workload communication in training/inference — RDMA, NCCL, topology-aware optimization, or high-performance networking is a strong plus
• Strong Linux, container runtime, and node-level systems fundamentals
• Proficiency in at least one of Go, Python, or C++
• Strong execution and cross-functional collaboration skills

Nice to Have

• Production experience operating clusters at 1,000+ GPU scale
• Experience with Volcano, Kueue, or Kubernetes scheduler extensions
• Familiarity with InfiniBand, RoCE, EFA, NVLink, or GPUDirect
• Experience with distributed storage, high-throughput data access, checkpoint management, or object storage governance
• Experience with observability stacks: Prometheus, Grafana, Loki/ELK, DCGM Exporter
• Experience with multi-cloud or hybrid-cloud AI infrastructure

AIM Global Talent Pte. Ltd. | EA Licence No. 25C3207