Confirmed on the employer's own hiring board on Nov 18, 2026. First seen by Alion on Nov 17, 2026.
Description
Location: Ra'anana
#HybridAbout DriveNets
DriveNets is a leader in large-scale networking solutions for AI infrastructure and service providers. Founded in 2016, the company created a radical new way to build networks that substantially improves the network’s economic model and optimizes its utilization and efficiency.
DriveNets Network Cloud powers production networks for global tier-1 operators such as AT&T, Comcast and KDDI, and carries the majority of AT&T’s core backbone traffic. DriveNets AI Fabric, introduced in 2023, provides the highest-performance Ethernet-based AI networking solutions and the best Ethernet alternative to InfiniBand. It is used by hyperscalers, NeoClouds, and foundation model labs, and is sold in collaboration with partners like AMD and Broadcom. In June 2026, DriveNets closed a $410 million Series D round, bringing its total capital raised to $1 billion. The company has more than $1 billion in secured business.
Join us to build the infrastructure of the largest AI Factories in the world and the networks behind the world’s leading telecom operators. We are looking for talented people with experience in networking architecture, AI infrastructure, and Cybersecurity, who want to make a real impact. Learn more at drivenets.com.
Responsibilities:
- Build, lead and mentor the DNOS-SAI SIT team; own the quality of DNOS-SAI releases on Broadcom Tomahawk (TH5/TH6) platforms.
- Learn, investigate and write test plans for AI data-center use cases: leaf-spine and rail-optimized Clos fabrics, lossless Ethernet (PFC/ECN/DCQCN), ECMP, eBGP underlay, EVPN/VXLAN and multi-tenancy.
- Simulate and validate the integration between routing, switching and QoS features in complex environments: BGP, EVPN/VXLAN, ECMP, QoS/PFC/ECN, LACP, BFD, LLDP, ACLs, telemetry, ZTP/upgrades, port breakout and 400G/800G/1.6T optics.
- Run testing scripts and traffic scenarios in the lab (Spirent/Ixia, RDMA and collective traffic) and analyze counters, drops, latency and congestion at the switch and at the GPU NIC.
- Validate GPU clusters and AI workloads end to end: RoCEv2/RDMA performance, NCCL/RCCL benchmarks and training/inference jobs under SLURM or Kubernetes.
- Lead defect investigation with R&D (reproduce, isolate ASIC/SAI vs DNOS behavior, verify fixes, run regression) and be an integrated part of network design and bring-up for our AI customers.
Requirements
- 3+ years leading a QA/SIT team on top of 5+ years of hands-on system/integration testing of networking products.
- Deep knowledge and experience in data-center routing and switching: BGP, EVPN/VXLAN, ECMP, VLANs, IP addressing, MTU and L2/L3 troubleshooting.
- AI networking: hands-on experience with AI/HPC fabrics - RoCEv2/RDMA, lossless Ethernet (PFC/ECN/DCQCN), rail-optimized Clos designs, 400G/800G/1.6T optics.
- GPU clusters and AI workloads: running and monitoring NCCL/RCCL, RDMA and MPI benchmarks and training/inference jobs on NVIDIA/AMD GPU servers (SLURM or Kubernetes).
- Knowledge and experience with Linux systems.
- Knowledge with services such as LLDP, LACP, BFD, SNMP, TACACS/RADIUS, QoS and gNMI telemetry.
- Hands-on use of AI agents in daily engineering work (skills, rules files, MCP tools; e.g., Claude Code, Codex, Cursor).
- Able to work in complex environment/tasks; self-learning and issue solving.
- Good communication skills, worldwide customer facing, good English.
Advantage:
- CCNP, CCIE certificate or equivalent knowledge.
- Experience with white-box switches and SAI/SONiC-based network operating systems on Broadcom Tomahawk ASICs.
- Service-provider technologies such as IS-IS/OSPF, MPLS/Segment Routing, L3VPN, SR-TE and multicast (PIM, MVPN).
- Testing tools such as Ixia or Spirent, and RDMA/collective tools (perftest, NCCL/RCCL tests, OSU benchmarks).
- Automation scripts in Python, Bash or Java and CI pipelines; Kubernetes, containers and cloud deployments (AWS/Azure/GCP).
- QA methodologies (HLD/LLD-based test plans) and tools such as Jira and TestRail.

