Senior AI Infrastructure Engineer — Network Deployment
- Compensation
- Not disclosed
- Location
- Bangalore, Karnataka, India
- Work setup
- Onsite
- Seniority
- Senior
- Employment
- Not specified
- Experience
- 5+ years
- Category
- Devops
- Posted
Skills
Must have
- 5–8 years in hands-on data center networking.
Full description
About Nava
Nava is a neocloud company purpose-built for the AI era. We design, deploy, and operate large-scale GPU infrastructure and deliver inference-as-a-service to teams building the next generation of AI products. Our platform runs on NVIDIA GPU systems, high-performance RDMA fabrics, and a fully automated, software-defined operations model. Every engineer at Nava works close to the metal, on infrastructure built to keep GPUs saturated and models serving.
About the Role
The Network Deployment team builds and validates the high-speed switching and GPU networking that connects Nava's production AI clusters. This is a hands-on engineering role.
Deploy and validate high-speed GPU cluster networks, including InfiniBand, RoCE, and Ethernet fabrics. Configure switches, automate provisioning, troubleshoot connectivity, benchmark network performance, and ensure production-ready infrastructure.
What You’ll Do
-
Configure InfiniBand/RoCE, Ethernet leaf-spine, storage and management networking.
-
Perform optics, cable, routing, topology and link-state validation.
-
Test RDMA reachability, bandwidth, latency and fabric congestion under load.
-
Automate switch provisioning, preflight tests and configuration checks.
-
Investigate hardware, routing and network fabric issues with compute/storage teams.
-
Produce acceptance evidence, as-built documentation and operational handover.
-
Configure and tune PFC/ECN for lossless RoCEv2 (or InfiniBand credit flow), validate rail-optimized topology and adaptive routing, and run NCCL bus-bandwidth acceptance tests.
-
Use UFM (InfiniBand) and NetQ / Spectrum-X telemetry to verify PFC-pause is near zero and GPUDirect RDMA / BlueField DPU paths perform at line rate.
What You’ll Need
-
5–8 years in hands-on data center networking.
-
Strong BGP, L2/L3, EVPN/VXLAN and high-speed switching knowledge.
-
Linux debugging and Python/Ansible automation.
-
Experience with Arista, Cisco, NVIDIA Networking or similar platforms.
-
Within that, hands-on GPU / AI-datacenter exposure is what we weigh most — we value demonstrated depth over years, as the field is still emerging.
- RDMA/RoCEv2, InfiniBand, PFC/ECN and GPU cluster testing.