Lead Principal Engineer, Enterprise Agentic AI Platform

NVIDIA•Santa Clara, CA

6d•$272,000 - $431,250

About The Position

Join NVIDIA IT’s Enterprise AI & Automation team to develop and expand enterprise-grade agentic AI systems. NVIDIA’s Enterprise AI Platform supports production AI agents that securely connect with enterprise systems to improve employee efficiency and business outcomes across engineering, IT, supply chain, finance, HR, and sales. This is a hands-on architecture role for a Principal or Distinguished Engineer–level technical leader who writes code daily in Python and/or Go, rapidly prototypes with tools such as Cursor, Claude Code, and Claude Cowork, and understands infrastructure ranging from Kubernetes to GPU inference stacks. The role will define NVIDIA’s enterprise agent architecture by building functional systems, developing reference implementations, and raising technical standards through production systems and measurable impact. It focuses on the full agent development lifecycle—creating, sandboxing, launching, observing, controlling, and continually improving agents through data-driven cycles. The systems will incorporate persistent memory, controlled runtime environments, rigorous assessment, and GPU-powered performance to make agents intelligent, trackable, protected, and production-ready.

Requirements

Bachelor’s degree in Computer Science or a related field, or equivalent experience.
15+ years of experience building and shipping large-scale distributed systems, with significant hands-on coding in Python, Go, or similar systems languages.
Proven ability to move quickly from an idea to a functional prototype and then to a robust, scalable platform solution.
Proven track record constructing agentic AI systems, including RAG pipelines, long-lasting memory models, multi-agent management (such as LangChain or LangGraph), tool frameworks, and evaluation infrastructure.
Expert-level depth in Kubernetes, containerized workloads, networking, APIs, and secure enterprise integration patterns.
Experience building benchmarking, regression testing, telemetry, and observability systems that measure agent quality, latency, cost, reliability, and safety.
Comprehensive knowledge of performance tuning in hybrid environments, including GPU-based inference systems.
Excellent collaboration skills, including influencing cross-functional collaborators, building positive relationships, and clearly communicating complex architectural concepts to technical and business audiences.

Nice To Haves

Master’s or PhD.
Experience delivering reusable developer-acceleration components such as SDKs, APIs, templates, reference implementations, and CI/CD automation.
Experience integrating enterprise vector databases and retrieval systems, and working with agentic search and orchestration platforms such as Glean, Microsoft Copilot Studio, Google Agentspace, or similar enterprise AI ecosystems.
Experience embedding fine-grained policy enforcement, access controls, sandbox isolation, and audit trails directly into AI runtimes.
GPU-acceleration experience optimizing model inference, batching strategies, memory utilization, and efficiency on NVIDIA hardware.
Meaningful open-source contributions, such as core commits, maintainership, widely adopted libraries, or public technical artifacts demonstrating system-level depth.

Responsibilities

Develop and deliver production-quality agentic AI systems end to end using Python and/or Go, including Kubernetes deployment, agent runtimes, memory systems, orchestration, tool integration, and evaluation pipelines.
Define and advance NVIDIA’s Enterprise Agentic AI architecture through practical implementations, reference systems, and production deployments.
Build and implement multi-agent orchestration patterns, such as planner, executor, reviewer, and tool agents, using frameworks such as LangChain, LangGraph, or similar systems, with regression coverage and observability.
Rapidly prototype emerging agent architectures and harden successful patterns into reusable platform services, APIs, SDKs, and developer templates.
Architect and implement data flywheels that improve agent quality through telemetry, benchmarking, automated evaluation, and structured feedback loops.
Embed security, guardrails, sandbox isolation, auditability, and policy enforcement into agent runtimes in partnership with security and governance teams.
Evaluate, integrate, and extend open-source and third-party agent platforms, making build-versus-use decisions based on performance, scalability, control, and long-term platform ownership.
Collaborate with engineering, infrastructure, product, and business collaborators to align architectural direction with enterprise priorities and accelerate adoption.