AI Software Engineer

NetApp, Inc.

19h•$170,000 - $253,000•Onsite

About The Position

As a Sr Engineer on the AI Enablement team, you will be a hands-on builder and a force multiplier for how the broader Keystone engineering org works with AI. You'll split your time between scaling engineering practice and shipping autonomous agents that do real work — not prototypes that stall after a demo. This role requires strong software engineering fundamentals, comfort operating with ambiguity on a fast-moving small team, and the judgment to know when an AI-generated answer is good enough to ship versus a production risk.

Requirements

Bachelor's degree in Computer Science, Engineering, or a related field; or equivalent, relevant experience.
8+ years of professional software development experience, with proven ownership of at least one system or domain end-to-end.
Production experience with LLMs - prompt engineering, agent development, and evaluation frameworks.
Hands-on experience building agent harnesses: tool calling, state/session management, retries, guardrails, and stop conditions for long-running agent workflows.
Deep, hands-on proficiency in Python/Go/Java
Experience with agent/orchestration frameworks (LangChain, LangGraph, or equivalent) and MCP or similar tool-calling standards, integrated with enterprise systems
Working knowledge of AI-native, spec-driven development workflows (agentic IDEs like Cursor/Claude Code) and defining quality gates for AI-generated code.
Strong problem decomposition and stakeholder navigation - able to scope an ambiguous cross-functional ask into a shippable plan and own it end-to-end, forward-deployed-engineer style.

Responsibilities

Scale engineering with AI
Drive spec-driven, agentic-IDE development workflows (Cursor, Claude Code) across the team, from spec to implementation to review, so AI usage translates into measurable velocity, not just novelty.
Define and evolve quality gates for AI-generated code: what changes in review, CI, and merge criteria when a large share of PRs are AI-assisted.
Build and maintain shared skills, prompts, and MCP tooling that other Keystone engineers reuse, reducing redundant AI infrastructure across teams.
Build production AI agents on the enterprise stack
Design and ship agent workflows (planner vs. fixed-workflow, state/retries, human-in-the-loop checkpoints) that automate real toil
Build well-scoped MCP tools/integrations against enterprise systems with clear schemas, auth boundaries, idempotency, and explicit rules on what data never reaches the model.
Own evals and guardrails for agents in production: offline golden sets, online sampling, canary/rollback paths, and hard checks on numeric/PII/outbound-send correctness — distinguishing tool failures from model failures.
Design for operational readiness from day one: what breaks under a 10× usage spike before quarter close, what the kill switch looks like when an agent gets a fact wrong, and how incidents get triaged.