Job Summary:
Moonlite AI delivers high-performance AI infrastructure for organizations running intensive computational research and data processing workloads. The Senior Software Engineer, Network Platform will be responsible for building a software-defined networking platform that enables high-performance, isolated networking for distributed computing and model training.
Responsibilities:
• Collaborate with infrastructure to design and build scalable SDN orchestration systems leveraging NVIDIA Bluefield-3 DPUs to deliver programmable, high-performance networking for AI workloads with hardware-accelerated forwarding isolation.
• Design and implement networking systems for research computing environments including Kubernetes and SLURM clusters, enabling high-performance connectivity, optimized network topology for distributed workloads, and seamless integration with cluster orchestration systems.
• Implement automated SDN provisioning systems that handle VPC creation, subnet allocation, routing configuration, and network resource lifecycle from deployment through decommissioning.
• Develop platform capabilities for managing Bluefield-3 DPUs including SR-IOV virtual function management, OVS offload configuration, network function deployment, and integration with compute orchestration systems.
• Build enterprise-grade network isolation using VPCs, VXLAN, and hardware-accelerated forwarding to ensure complete tenant separation while maintaining high-performance connectivity for GPU clusters and distributed workloads.
• Collaborate with infrastructure to optimize network paths for RDMA, RoCE, and GPU-to-GPU communication, ensuring minimal latency and maximum throughput for distributed training and large-scale computational workloads.
• Develop robust APIs and SDKs for network resource management that integrate seamlessly with compute and storage platforms, enabling programmatic network provisioning and configuration.
• Implement comprehensive network monitoring, telemetry, and troubleshooting systems that provide visibility into network performance, utilization, and tenant traffic patterns.
• Build platform network security features including security groups, firewall rules, and policy enforcement that protect tenant workloads while enabling flexible network configuration.
Qualifications:
Required:
• 5+ years in software engineering with proven experience building network platforms, SDN systems, or network automation for production environments.
• Strong familiarity with Kubernetes networking architecture, CNI plugins, service networking, and network policies. Understanding of pod networking, services, ingress, and how Kubernetes manages network resources.
• Deep understanding of networking fundamentals including TCP/IP, VLANs, VXLAN, BGP, OSPF, routing protocols, and data center network architectures.
• Background in SDN concepts, network virtualization, overlay networks, and programmable networking technologies.
• Experience with Go and Python for performance-critical networking components and services is highly valued.
• Strong experience with Linux networking stack, including network namespaces, iptables/nftables, Open vSwitch, and kernel networking systems.
• Familiarity with DPU/SmartNIC architectures (Bluefield, or similar), SR-IOV, hardware offload capabilities, and programmable networking hardware – or strong ability to learn quickly.
• Understanding of RDMA, RoCE, Infiniband, and low-latency networking requirements for distributed computing and GPU workloads.
• Demonstrated ability to solve complex networking performance and scalability challenges while balancing pragmatic shipping with good long-term architecture.
• Comfortable navigating ambiguity, defining requirements collaboratively, and communicating technical decisions through clear documentation.
• Growth mindset with continuous focus on learning and professional development.
Preferred:
• Background provisioning or managing networking for research computing environments (Kubernetes, SLURM, or HPC clusters)
• Experience with NVIDIA Bluefield DPU programming and DOCA framework
• Background with network function virtualization (NFV) and service function chaining
• Knowledge of Kubernetes networking (CNI plugins, network policies, service mesh)
• Experience building network control planes or SDN controllers
• Familiarity with network automation frameworks and infrastructure-as-code for networking
• Understanding of data center fabric architectures (spine-leaf, CLOS topologies)
• Experience with network security and compliance requirements in regulated industries
• Background building networking for research institutions, HPC environments, or cloud providers
Company:
Moonlite is building a cloud-native experience on-prem. Our software provides the control and customization enterprises need for AI. Founded in 2024, the company is headquartered in Chicago, US, , with a team of 2-10 employees. The company is currently Early Stage.