Senior AI/HPC Network Architect

Groq, Inc.; New York (NY); Hybrid; USD 270,000 - 402,000; Full time

Job summary

Groq is seeking a Sr./Staff Network Design Engineer to architect and validate the network infrastructure for AI compute environments. Based in the Dallas, San Francisco, or New York City area, you will own end-to-end designs, collaborate with software teams, and drive automation and observability at scale.

You will build spine-leaf fabrics, evaluate EVPN/VXLAN, BGP, QoS, and interconnect strategies, while mentoring others and documenting standards for production deployments.

Qualifications

• 8+ years of network engineering, architecture or infra engineering experience.
• Strong TCP/IP, BGP and data center routing knowledge.
• Experience designing leaf-spine/Clos fabrics.
• EVPN/VXLAN and modern network virtualization knowledge.
• Understanding ECMP, routing convergence, traffic engineering and resiliency.
• Experience with high-bandwidth Ethernet (100G/400G).
• Familiar with AI/ML clusters, HPC, or large GPU deployments.
• Proficiency in Python and APIs for automation; IaC practices.

Responsibilities

• Own end-to-end network design for AI training and inference.
• Collaborate with software teams to align topology with stack architecture.
• Develop high-level and low-level designs, BOMs, and standards.
• Design spine-leaf fabrics with performance, resiliency and scale in mind.
• Evaluate BGP/EVPN, VXLAN, IPv4/IPv6, QoS, and related tech.
• Plan data center interconnect, WAN and cloud connectivity.
• Perform capacity planning and traffic engineering for AI workloads.
• Translate workload patterns into network requirements.
• Establish standards for topology, addressing and routing policies.
• Review hardware platforms, OS, optics and cabling options.
• Build labs, proofs of concept and validation frameworks.
• Define failure scenarios and validate behavior under outages.
• Drive network automation using Python, Ansible, Git, APIs, CI/CD.
• Create automated validation, configuration generation and testing.
• Define observability: telemetry, metrics, alerts and capacity visibility.
• Lead design reviews and mentor engineers on architecture and tooling.
• Troubleshoot issues spanning network, compute, hardware and distributed systems.
• Partner with ops to ensure designs are observable and safe at scale.

Skills

TCP/IP

BGP

Leaf-spine / Clos fabrics

EVPN/VXLAN

ECMP

Network automation

Python

APIs

Git

CI/CD

Linux networking

Telemetry/Observability

Tools

Arista EOS

Cisco IOS

Juniper Junos

NVIDIA Networking

Job description

Groq is seeking a Sr./Staff Network Design Engineer to architect and validate the network infrastructure for AI compute environments. Based in the Dallas, San Francisco, or New York City area, you will own end-to-end designs, collaborate with software teams, and drive automation and observability at scale.

You will build spine-leaf fabrics, evaluate EVPN/VXLAN, BGP, QoS, and interconnect strategies, while mentoring others and documenting standards for production deployments.