📋
Description
- Together AI is building the next-generation AI compute platform, with networking at its core.
- As a Network Architect, you will define and evolve the global network architecture powering AI training, inference, and research platforms.
- You will own routing, topology, traffic engineering, and control-plane strategies across data centers, cloud environments, and backbone fabrics.
- The role involves designing high-bandwidth, low-latency, fault-tolerant networks supporting large-scale GPU clusters and HPC-style compute fabrics.
- Responsibilities include establishing topology strategies, traffic engineering, multicloud interconnects, control-plane architecture, observability primitives, and guiding emerging technologies.
- You will work closely with infrastructure, compute, storage, hardware, and operations teams to meet performance demands, support distributed workloads, and ensure supportability and resilience.
- Mentorship and strategic influence across the organization are key aspects of this role.
🎯
Requirements
- Deep experience designing and operating large-scale GPU clusters or HPC-style compute fabrics, understanding their network demands.
- Fluent in building high-throughput data center fabrics supporting tens of thousands of GPUs and multi-terabit east–west traffic.
- Experience designing or operating RoCEv2 or lossless Ethernet environments at scale, including PFC/ECN tuning and congestion control.
- Experience designing backbone and DCI architectures supporting GPU training clusters across multiple regions and fabrics.
- Led network architecture for multi-cloud, private backbone, and diverse PoP environments, understanding AI workload behaviors.
- Design experience with operational considerations: observability, capacity modeling, automation, telemetry, failure analysis.
- Ability to set architectural direction in fast-evolving compute, storage, and network environments.
🏖️
Benefits
- Competitive compensation
- Startup equity
- Health insurance
- Other competitive benefits