Senior Software Engineer, Agentic AI

NVIDIA•Austin, TX

21h

About The Position

NVIDIA is seeking an experienced Software Engineer to develop core libraries for Agentic Applications. This role involves building foundational technology for the next generation of agents, focusing on enabling the safe and scalable deployment of autonomous systems. As a leader in AI and deep learning, NVIDIA is at the forefront of agentic computing. The successful candidate will contribute to building scalable agentic capabilities, reusable building blocks, and high-quality libraries to accelerate developer productivity, ensure agent quality, and optimize agent performance and efficiency. This position offers the opportunity to collaborate with creative engineers on projects that push the boundaries of autonomous agents.

Requirements

Bachelor's or Master's degree in Computer Science, Artificial Intelligence, Applied Math, or a related field, or equivalent experience.
8+ years experience in at least one of Rust, Python, Go, or Node.js, with working familiarity in at least one more.
Solid understanding of asynchronous programming, callbacks, request lifecycles, and event-driven systems.
Hands-on experience with evolving agent architectures, multiple agents frameworks and agent harnesses.
Proficiency in LLM applications, agent workflows, tool calls, and model-provider APIs.
Ability to design or extend cross-language APIs with attention to consistency, usability, stability, and backwards compatibility.
Systems-level debugging and performance intuition: able to trace execution from high-level API calls through runtime internals, language bindings, callbacks, serialization, and event emission to understand overhead and optimize hot paths.
Strong interpersonal skills, enabling direct contact and effective collaboration with the open source community.

Nice To Haves

Experience building evaluation/benchmarking systems for agent workflows (metrics, regression, feedback loops).
Rust systems work, especially async Rust, Tokio, serde, API design, or runtime state management and /or Python native extension experience with PyO3, maturin, or Python/Rust bindings.
Instrumenting third-party frameworks without changing user-visible behavior.
Knowledge of OpenTelemetry, tracing, structured events, exporters, or observability pipelines are a plus.
Middleware, plugin systems, guardrails, policy engines, or request/response interception and maintaining open-source libraries, SDKs, or internal developer platforms.
Profiling or optimizing runtime/library overhead across language boundaries, async execution, native bindings, serialization, tracing, or middleware pipelines.

Responsibilities

Track and understand evolving agent development patterns across NVIDIA and the broader ecosystem, maintaining current knowledge of both research and commercial products.
Develop open-source libraries and tools which accelerate and optimize agent harnesses and frameworks ensuring top-tier performance, accuracy, quality, and stability.
Benchmark the latest agents to identify bottlenecks and build creative solutions to increase performance, reduce cost, and improve latency.
Work closely with teams building high-performance data pipelines, RAG systems, vector databases, and GPU-optimized training and inference workflows to deliver best-in-class agentic applications.
Identify gaps and friction in current agent architectures, and translate insights into agentic tools that boosts developer velocity and agent quality—backed by evaluations, benchmarking, and feedback loops.