Network Engineer - AI Infrastructure, RDMA & Storage

VAST Data, Inc.; Singapore; In loco; SGD 100.000 - 150.000; Tempo pieno

Descrizione del lavoro

VAST Data is seeking a Network Engineer—AI Infrastructure, focusing on RDMA and storage—to design architectures, run a dedicated lab, and validate networking gear for enterprise AI deployments in Singapore. You will collaborate with R&D, QA, and vendor teams to ensure high-speed, lossless, scalable networks for GPU fabrics and storage protocols.

Responsibilities include architecting L2/L3 designs, validating switches and NICs, and reproducing customer issues in a lab, turning findings into

Competenze

• Strong data center networking fundamentals — CCNP level or equivalent hands-on experience.
• Solid command of BGP and L3 routing: eBGP/iBGP, ECMP, route policy, multipath, BGP-based fabrics.
• Hands-on experience designing and troubleshooting both L2 and L3 data center networks.
• Real troubleshooting ability across layers—routing, ARP/ND, MTU, congestion, packet loss, latency.
• Working knowledge of Linux networking: interfaces, VLANs, routing tables, namespaces, MTU.
• Clear technical writing—designs and findings should be easily followable.

Mansioni

• Design reference network architectures for deployments: L2/L3 fabrics, BGP/EVPN/VXLAN, RoCEv2, QoS.
• Own the lab: build and run validation environments for switch qualification and reproduction of issues.
• Qualify the ecosystem: evaluate switches/OS and NICs from major vendors and certify compatibility.
• Solve complex problems across the path between GPUs, fabric, and storage; diagnose root causes.
• Turn findings into product: contribute to best practices and input to R&D.

Conoscenze

BGP,L3 routing

L2,L3 data center networks

Network lab design

Troubleshooting

Lab reproduction of problems

Technical writing

Strumenti

tcpdump

Wireshark

iperf

Descrizione del lavoro

Network Engineer - AI Infrastructure, RDMA & Storage

• Engineering
• Singapore
• Full-time

Description

VAST Data is the data platform company for the AI era. We build enterprise software infrastructure to capture, catalog, refine, enrich, and protect massive datasets, and make them available for real-time data analysis and AI training and inference. Designed from the ground up to make AI simple to deploy and manage, VAST takes the cost and complexity out of deploying enterprise and AI infrastructure across data center, edge, and cloud.

About the Role :

The fastest system in the world is only as fast as the network under it. As a Network Engineer in VAST's Field Engineering organization, you define how VAST connects to the world: the reference architectures, validated switch configurations, and RDMA fabric designs every customer deployment is built on.

This is a design and research role, not a support queue. You will build and run the lab where new switches, NICs, and network operating systems are qualified before they ever reach a customer. You will investigate why a fabric behaves differently at 800G than it did at 100G, and turn the answer into a standard the whole field organization uses. When a strategic customer hits something nobody has seen before, you reproduce it, explain it, and make sure it never happens to the next customer.

You will work across R&D, QA, Support, Product, and customer engineering teams, and directly with networking vendors including NVIDIA, Arista, Cisco, and Juniper. The work sits where AI infrastructure is hardest: GPU fabrics, lossless Ethernet, and storage protocols moving at line rate.

Requirements

What a Day in the Life Looks Like :

Design the Architecture:

Define reference network architectures for VAST deployments: L2 and L3 fabrics, BGP and EVPN/VXLAN designs, RoCEv2 and QoS policy. Decide what "good" looks like, then write it down so the field can execute it consistently.

Own the Lab:

Build and run validation environments for switch qualification, interoperability testing, performance characterization, and reproduction of customer issues. Your lab is where we find problems before customers do.

Qualify the Ecosystem:

Evaluate new switches, network operating systems, and NICs from NVIDIA, Arista, Cisco, Juniper, and others. Work directly with vendor engineering teams to close gaps and certify what works.

Solve What Nobody Else Can:

Chase complex problems across the full path between GPUs, fabric, and storage. Congestion, packet loss, MTU, latency, PFC and ECN behavior under real load. Root cause it, then fix it at the source.

Turn Findings into Product:

Feed what you learn back into VAST, as documented best practice, as automation, and as direct input to R&D on what the platform should do differently.

Professional Experience & Skills

• Strong data center networking fundamentals - CCNP level or equivalent hands-on experience
• Solid command of BGP and L3 routing: eBGP/iBGP, ECMP, route policy, multipath, BGP-based fabrics
• Hands-on experience designing and troubleshooting both L2 and L3 data center networks
• Real troubleshooting ability across layers - routing, ARP/ND, MTU, congestion, packet loss, latency
• Working knowledge of Linux networking: interfaces, VLANs, routing tables, namespaces, MTU
• Comfort with tcpdump, Wireshark, iperf, and the standard diagnostic toolkit
• The instinct to build a lab and reproduce a problem rather than theorize about it
• Clear technical writing - your designs and findings are only useful if others can follow them