NVIDIA is looking for an exceptional engineer to grow and thrive alongside our Infrastructure team that designs, evaluates, and optimizes the compute inrastructure powering NVIDIA's next-generation silicon design and AI
workloads.You will own the operational health, efficiency and observability of a large Linux compute farm serving NVIDIA's VLSI design teams
What you'll be doing
In this position, you will own the operational health, efficiency and observability of a large Linux compute farm serving NVIDIA's VLSI design teams, and be the person who can say where the time and the capacity are actually going.
Profile and debug at the system level - kernel, filesystem, memory and I/O behaviour - when workloads are slow, and turn findings into fixes rather than tickets.
Drive compute efficiency: memory request-versus-usage accuracy, CPU waste from idle and over-provisioned jobs, and the prediction models behind resource requests.
Build the observability layer for the farm - telemetry pipelines, metrics and dashboards that make capacity, utilisation and job behaviour visible to engineers and to leadership - and own licence monitoring and expiry alerting across the EDA tool estate.
This is a hands-on technical role. Day-to-day farm operations and live triage are a core part of it, and you will work directly with the heaviest users on their workload behaviour, since most efficiency gains come from a small number of workflows.
Show the full description
What we need to see
B.E./B.Tech/M.Tech in Computer Science, Electrical Engineering, or equivalent experience.
8+ years in Linux systems administration or infrastructure operations at scale.
Deep Linux systems and kernel-level expertise - comfortable with perf, eBPF/bpftrace, strace, cgroups, NUMA, memory management, scheduler behaviour and I/O subsystems, and able to use them to find root cause rather than guess at it.
Demonstrated experience diagnosing and fixing performance problems on large multi-user compute systems.
Strong NFS and distributed filesystem troubleshooting: metadata versus data path, client behaviour, mount and latency issues.
Hands-on experience building observability - time-series metrics, log and telemetry pipelines, and dashboards (Prometheus/Grafana, Elastic, or equivalent) - plus strong Python for analysis and automation.
Familiarity with batch schedulers (LSF, Slurm, or equivalent) and job resource accounting, and a genuine appetite for operational work. Semiconductor/EDA or HPC background is a plus.
