Senior Software Engineer - Backend Platform

NVIDIA Corporation; United States; On-site; USD 180,000 - 240,000; Full time

Job summary

NVIDIA is seeking a Senior Software Engineer to build the backend platform for its Observability and Prediction stack, processing telemetry from GPUs, NICs, switches and links at extreme scale. You'll design and implement distributed services, storage models, and resilient deployment pipelines across clusters from a single node to thousands of GPUs.

You'll work in a polyglot environment (Go, C++, Rust, Python), own throughput, latency, and back-pressure, and collaborate with teams to ensure

Qualifications

• B.Sc./M.Sc. in Computer Science, Computer Engineering, or a related technical field.
• 5+ years of software engineering experience building production backend systems.

Responsibilities

• Build the high-scale data path: ingest, normalize, enrich and route telemetry streams with throughput, concurrency and horizontal scalability.
• Design the storage and query layer for large datasets with fast dashboards, alerts and analyses.
• Make it resilient: handle partial failure, lost state, and restarts without silent failures.
• Own platform shipping and deployment: packaging, upgrades, backups, and on-prem/off-cloud deployments.

Tools

Go

C++

Rust

Python

Job description

NVIDIA is powering the world's most advanced AI Factories. Keeping them running depends on an Observability and Prediction platform that continuously ingests telemetry from every GPU, NIC, switch and link in the fleet - delivered both as a high-scale SaaS service and as a self-contained on-premises deployment for our largest enterprise customers.

We are looking for a Senior Software Engineer to build the backend platform underneath it: the distributed services that move, transform, store and serve telemetry at extreme volume, and the infrastructure that makes them deployable, upgradable and survivable on clusters ranging from a single node to tens of thousands of GPUs. This is the substrate every analytics, alerting and AI layer runs on - if it isn't fast, correct and resilient, nothing above it works.

What you'll be doing:

• Build the high-scale data path. Design and implement the services that ingest, normalize, enrich and route enormous, continuous telemetry streams. You'll own throughput, back-pressure, concurrency, and how the system scales horizontally as clusters grow by an order of magnitude.
• Design the storage and query layer. Choose the right data model and store for each kind of data, and make queries over very large datasets return fast enough to drive live dashboards, alerting and automated analysis. Own the trade-offs - cardinality, retention, cost, latency - and the APIs the rest of the product is built on.
• Make it resilient. Design for partial failure: what happens when a node disappears, when state is lost, when a service restarts into the wrong configuration. Eliminate silent-failure modes where a service looks healthy while delivering nothing.
• Own how the platform ships and runs. Packaging, configuration, upgrade paths, backup and restore, and clean deployment onto customer-controlled environments - so that software you can't log into still behaves correctly on day 400.
• Push on performance and capacity. Profile the system under realistic load, find the real bottleneck rather than the suspected one, fix it, and translate the results into concrete sizing and scaling guidance.

What we need to see:

• B.Sc./M.Sc. in Computer Science, Computer Engineering, or a related technical field.
• 5+ years of software engineering experience building production backend systems.
• Strong proficiency in languages as Go, C++, Rust or Python, with solid fundamentals in concurrency, memory and performance - and the flexibility to work across languages in a polyglot codebase.
• Hands-on experience with distributed systems: services that talk to each other over a network, fail independently, and have to stay correct anyway.
• Practical experience with at least one class of data infrastructure - streaming/messaging systems, databases (relational, time-series, analytical, graph or key-value), or high-throughput data pipelines. We care that you understand the trade-offs, not that you've used our specific stack.

Ways to stand out from the crowd:

• Experience shipping software that other organizations install and operate themselves: versioning, compatibility, upgrades, and debugging environments you don't control.
• Depth in large-scale data systems - you've dealt with the point where the obvious design stops working and had to redesign it.
• Networking or datacenter-infrastructure background, or experience with telemetry, monitoring and observability systems.
• A systems thinker: you understand the full stack, from how data moves across the wire to how it's processed in a distributed cluster, and you can tell us where the time went.