Senior/Staff Software Engineer, AI Agent Infrastructure

Nuro•Mountain View, CA
•$193,930 - $352,290

About The Position

Nuro is hiring a Senior/Staff Software Engineer for its AI Agent Infrastructure team. The team builds infrastructure that enables AI agents to operate autonomously within Nuro’s engineering organization, with rigorous evaluation and evidence-based safeguards. The role will help make Nuro’s production agent system more trustworthy, autonomous, and capable. The position focuses on three areas: closed-loop evaluation that measures agent outcomes and informs whether agents can be trusted to act; an agent platform with orchestration, sandboxing, isolation, tools and skills, memory, identity and permissioning, gateways, and observability; and autoresearch infrastructure that automates model research experiments, evaluation, and proposals. The team also develops agent-powered tools across the engineering lifecycle and may post-train internal models where workloads justify it. The engineer will work end-to-end with substantial ownership, directly alongside engineering leadership and the CEO. Potential early projects include building workflow-level measurement of agent acceptance, reverts, and overrides; taking a bounded class of autoresearch experiments to unattended operation; and designing auditable isolation and permissioning for agents working with production repositories and infrastructure.

Requirements

  • At least 5 years of software engineering experience, or at least 4 years with a Master's degree, in computer science, engineering, or equivalent practical experience.
  • Staff-level candidates must bring correspondingly deeper scope and ownership.
  • Deep, current understanding of LLM research and model training, including data, tokenization, architecture, pretraining, supervised fine-tuning, preference optimization, and reinforcement learning.
  • Understanding of inference internals, including attention, KV-cache behavior, batching and scheduling, quantization, speculative decoding, prefix caching, and context handling, and their effects on latency, throughput, and cost.
  • Experience building and operating production LLM-based agent systems, including tool use, orchestration, sandboxing, retrieval, and memory.
  • Strong backend and distributed systems experience at scale, including cloud infrastructure, service design, storage, and queuing.
  • Strong programming skills in Python.
  • Strong judgment about evaluation and whether benchmarks measure meaningful outcomes.
  • Ability to work end-to-end and take ownership without requiring a fully decomposed specification.
  • Hands-on post-training or fine-tuning experience, such as SFT, preference optimization, RL, or distillation, including evaluation to determine whether it helped.
  • Experience with ML training or research infrastructure, such as experiment orchestration, evaluation pipelines, hyperparameter search, or data pipelines.
  • Experience running inference serving, cost, or capacity at meaningful scale.

Nice To Haves

  • Familiarity with agent architecture patterns such as planning, reflection, long-horizon memory, and multi-agent coordination.
  • Experience with open tool-integration protocols, plugin or skill frameworks, and model-routing or gateway layers.
  • Background in developer experience or platform engineering, observability, or security isolation.
  • Go, C++, or Rust experience in addition to Python.

Responsibilities

  • Build closed-loop evaluation and measurement for agent workflows, including tracking acceptance, reverts, and overrides and detecting regressions.
  • Use evaluation results to determine where agent autonomy should expand or be constrained.
  • Develop agent runtime infrastructure, including orchestration, sandboxing, isolation, tools and skills, memory, identity and permissions, gateways, and observability.
  • Automate autoresearch workflows that form hypotheses, launch experiments, evaluate results, and propose or discard changes.
  • Build evaluation and confidence mechanisms that allow bounded research experiments to run unattended.
  • Design auditable isolation and permissioning for agents acting on production repositories and infrastructure.
  • Develop agent-powered tooling across the engineering lifecycle, including code generation, review, debugging, test and CI failure attribution, knowledge retrieval, and triage.
  • Contribute to post-training internal models where justified by internal workloads.

Benefits

  • Annual performance bonus
  • Equity
  • Competitive benefits package
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service