AI Infrastructure Engineer
Bitdeer is a NASDAQ-listed company specializing in Bitcoin mining, AI cloud services, and data center operations. They provide comprehensive mining solutions, AI development services, and operate seven global data centers with a significant portion of their energy being carbon-free.
About Bitdeer Technologies Group
Bitdeer is a high-performance computing platform that specializes in Bitcoin mining and AI cloud services. The company constructs and operates numerous cutting-edge data centers globally, including one of North America's largest, with a focus on using clean energy. Their product offerings include SEALMINER, which are high-performance and efficient chips; Cloud Mining services that allow users to participate in mining remotely; and Minerbase, a mobile intelligent cooling system. In addition to their crypto-focused services, Bitdeer is expanding into the AI sector by providing high-performance computing for AI/ML acceleration using NVIDIA DGX H100s, offering turnkey AI datacenter solutions, and developing an end-to-end AI application sharing ecosystem. They cater to both individual and institutional clients with a range of services like after-sales support, a hash rate market, and data center support.
Skills
About the Role
You will operate and optimize GPU clusters spanning multiple regions, working with Kubernetes, Slurm, and Ray to keep training and inference workloads running smoothly. You'll implement elastic scheduling and unified orchestration, balancing capacity between training and serving jobs. You will manage and tune high-throughput, low-latency inference runtimes, own model hot-swapping and zero-downtime rollout processes, and benchmark performance across a range of model sizes and workloads. You'll also tune distributed communication stacks and build observability tooling so the team can monitor GPU utilization, latency, and anomalies in real time.
Requirements
- Bachelor's or Master's degree in Computer Science, Electrical Engineering, or related field; PhD preferred for advanced R&D or innovation-oriented roles
- 3-5+ years in ML Infrastructure, HPC, or Systems Engineering
- Hands-on experience with Kubernetes, Slurm, or Ray
- Familiarity with vLLM, SGLang, or similar inference frameworks
- Strong background in PyTorch / JAX, distributed systems, and communication stacks (NCCL / RCCL, RDMA)
- Proficiency in Python plus one of Go / C++ / Rust
- Experience building observability with Prometheus and Grafana
- Fluent in English; experience working in multinational or cross-cultural environments is a plus
- Experience with major cloud platforms is strongly preferred
Responsibilities
- Operate and optimize GPU clusters using Kubernetes, Slurm, and Ray across multiple regions
- Implement elastic scheduling and unified orchestration for inference and training jobs, including preemption and dynamic capacity arbitration
- Manage and tune vLLM / SGLang runtimes for high-throughput, low-latency serving
- Optimize distributed scheduling for multi-replica, multi-tenant serving
- Own model hot-swapping and zero-downtime rollout paths
- Benchmark and profile performance across workloads and model sizes
- Tune distributed communication stacks including NCCL / RCCL, RDMA over RoCEv2, and InfiniBand
- Build observability with Prometheus, Grafana, and Ray Dashboard to monitor GPU utilization and latency
- Integrate observability with the platform-wide OpenTelemetry + Grafana LGTM+ stack
Benefits
- Attractive welfare benefits and developmental opportunities such as training and mentoring