Technical Leader, DPU Networking and AI Infrastructure

Cisco is building a next-generation DPU‑enabled networking platform for secure, high‑performance infrastructure and large‑scale AI clusters. Our team is responsible for designing the foundational networking, security, and programmable services that span host systems, DPUs, Smart

NICs, and cloud‑scale control planes.

This is a Hybrid position located in Milpitas, Ca. Ideally 3 days per week in office.

• Lead architecture for DPU‑accelerated host networking, security, and network services.
• Define what runs on the DPU versus the host CPU, including tradeoffs around performance, cost, failure isolation, observability, debuggability, and operational complexity.
• Architect distributed firewalling, micro segmentation, virtual routing, service chaining, traffic steering, and line‑rate policy enforcement.
• Guide platform decisions across DPU SDKs, programmable pipelines, Linux networking, control‑plane services, telemetry, policy distribution, and upgrade models.
• Design APIs, policy models, configuration flows, state synchronization mechanisms, and recovery behavior for distributed networking systems.
• Build static‑stability and failure‑mode thinking into the architecture so the system behaves predictably during partial failures, control‑plane loss, upgrades, rollbacks, and fleet‑wide rollout.
• Provide technical leadership for AI cluster networking challenges, including congestion collapse, incast, ECMP imbalance, PFC/ECN behavior, noisy tenants, NIC firmware issues, switch misconfiguration, and GPU under utilization caused by network stalls.
• Align cross‑functional stakeholders across hardware, firmware, product, and customer‑facing teams to successfully translate complex architectural visions into shippable software.
• Elevate engineering standards across the organization through meticulous design reviews, proactively identifying systemic risks to scale, reliability, and security.
• Mentor senior engineering talent, accelerating team growth and technical depth in packet processing, host networking, and control‑plane design.

• Bachelor’s degree with significant related experience, or advanced degree with equivalent experience in networking, systems software, distributed systems, or infrastructure engineering.
• Deep experience designing and building production networking systems, security platforms, cloud networking, host networking, or large‑scale infrastructure software.
• Proficiency in systems programming experience using C/C++, Go, Rust, Python, or similar languages.
• Applied skills in Linux networking, packet processing, routing, policy enforcement, virtualization, and production debugging.
• Demonstrated ability to lead architecture across multiple teams and drive complex technical decisions from concept through production.

• Hands‑on experience with DPU, Smart
  
  NIC, NIC offload, NPU, ASIC, or hardware‑accelerated networking platforms.
• Experience with NVIDIA Blue Field/DOCA, AMD Pensando, DPU SDKs, P4 or P4‑like programmable pipelines, DPDK, OVS, tc, eBPF/XDP, SR‑IOV, VF/PF, or kernel bypass.
• Experience architecting firewall, microsegmentation, NAT, virtual routing, VTEP, load balancing, traffic steering, service insertion, or service chaining systems.
• Proficient understanding of AI/HPC cluster networking behavior, including RDMA/RoCE, PFC, ECN, ECMP, congestion management, telemetry, and failure diagnosis.
• Experience designing distributed control planes with attention to consistency, state ownership, upgrade safety, rollback, policy convergence, and fleet operations.
• Ability to balance hardware offload, software portability, customer‑visible simplicity, performance, and long‑term platform maintainability.