Senior Software Engineer, Agentic AI

NVIDIA NVIDIA · Semiconductors · Redmond, WA +1

Senior Software Engineer role focused on building and scaling agentic AI systems for high-performance code generation. Responsibilities include architecting agentic systems, scaling distributed systems, developing evaluation frameworks, optimizing for performance on NVIDIA GPUs, and establishing engineering standards. Requires experience in building coding agents, AI evaluation, and distributed systems.

What you'd actually do

  1. Architect Agentic Systems: Lead the design and implementation of sophisticated coding agents and agentic platforms capable of multi-step reasoning, self-correction, and autonomous code generation.
  2. Scale Distributed Systems: Build and maintain high-performance, large-scale production systems that handle the compute-intensive nature of agentic workflows and AI evaluation.
  3. Develop Evaluation Frameworks: Create and implement rigorous AI evaluation methodologies to measure the accuracy, safety, and efficiency of generated code and agentic logic.
  4. Optimize for Performance: Collaborate with hardware and software teams to ensure agent-generated code and the underlying platforms are optimized for the latest NVIDIA GPU architectures.
  5. Establish Engineering Standards: Drive best practices for code quality, testing, and deployment within the agentic AI ecosystem to ensure reliable, production-grade output.

Skills

Required

  • MS in Computer Science, Computer Engineering, or related field (or equivalent experience)
  • 3+ years of experience building and maintaining large-scale production systems
  • Proven experience in building coding agents or developing comprehensive agent platforms
  • Direct experience developing skills and workflows using state-of-the-art coding agents and IDEs (e.g., Claude Code, Cursor, or Codex)
  • Proficient in Python
  • Proficient in at least one systems programming language (C++ or Rust)
  • Strong background in AI evaluation
  • Experience with distributed systems and orchestration

Nice to have

  • Experience writing CUDA kernels
  • Deep-dive profiling using Nsight Compute and Nsight Systems
  • Hands-on experience optimizing and deploying with TRT-LLM, SGLang, vLLM, or Transformer Engine
  • Demonstrated leadership in building and scaling agentic AI applications in production environments

What the JD emphasized

  • building coding agents
  • developing comprehensive agent platforms
  • AI evaluation
  • design benchmarks for complex, non-deterministic agentic behaviors
  • agentic AI applications in production environments

Other signals

  • building autonomous systems
  • agentic workflows
  • AI evaluation
  • distributed systems
  • NVIDIA’s world-class accelerated computing stack