Software Engineer, Applied AI Infrastructure

Highlights

• Build the most rigorous closed-loop evaluation system for AI work anywhere.
• Amplify every Nuro engineer's output 100x with trustworthy autonomous agents.
• Evolve a live agent fleet integrated into engineers' daily tools and workflows.

What they offer

• Annual performance bonus
• Equity
• Competitive benefits package

Ideal candidate

A rigorous, end-to-end builder obsessed with whether AI work can be trusted and measured. You thrive in ambiguity and want to amplify engineers 100x with autonomous systems held to real-world proof.

Requirements

• Care deeply about evaluation and whether benchmarks measure anything real
• Work end-to-end without needing problems pre-decomposed
• Hands-on post-training or fine-tuning experience including SFT, preference optimization, RL, distillation
• Experience with ML training or research infrastructure including experiment orchestration and evaluation pipelines
• Experience running inference serving, cost, or capacity at meaningful scale
• Familiarity with agent architecture patterns such as planning, reflection, memory, and multi-agent coordination
• Experience with tool-integration protocols, plugin or skill frameworks, and model-routing or gateway layers
• Background in developer experience, platform engineering, observability, or security isolation
• Experience with Go, C++, or Rust in addition to Python