Principal Software Engineer - Agentic AI

We're building an autonomous infrastructure engineering agent: think Claude Code, but for your cloud. If you're excited about agentic AI systems that solve impactful problems for engineers, developers, and security experts at innovative startups and iconic enterprises alike, this role is for you. Neo is Pulumi's answer to what infrastructure engineering looks like in an AI-first era. It's a virtual infrastructure teammate that understands your cloud environment, plans and executes multi-step operations, and brings domain expertise to problems ranging from spinning up complex ML infrastructures to cloud migrations to continuous compliance enforcement. What makes Neo possible is infrastructure-as-code: by modeling cloud infrastructure in code space, we can leverage the fact that LLMs are already excellent at coding to affect cloud infrastructure changes. Combined with Pulumi's registry and petabytes of deployment history, we have a unique foundation for building infrastructure intelligence. We're looking for engineers to build all aspects of Neo: from agent orchestration and tool interfaces to the web experience where users watch Neo work, approve operations, and understand its reasoning, to the interface with the models and how we orchestrate and fine-tune them to maximize impact. We seek bold thinkers to reimagine the entirety of cloud engineering. While Pulumi is headquartered in Seattle, WA., we are dedicated to being a remote first company. This role will be located remotely. Responsibilities Build the planning and execution systems that let Neo safely perform multi-step infrastructure operations across AWS, Azure, GCP, Kubernetes, and more Design tool interfaces that give the agent semantic understanding of cloud resources and access to engineering workflows Develop the context systems that help Neo understand each customer's infrastructure patterns and usage Create the feedback loops and self-correction mechanisms that make autonomous operation reliable Build evaluation frameworks to measure agent performance across Neo's capability areas: migrations, compliance, cost optimization, diagnostics, and more Work on the web interface where Neo operates as a "virtual teammate"—task assignment, progress visibility, approval workflows Integrate Neo's capabilities into Pulumi's MCP Server for agent-to-agent workflows Collaborate with platform teams to deeply integrate agent capabilities into the CLI, SDKs, and Pulumi Cloud You may be a good fit if you: Are genuinely excited about agentic AI—systems that plan, act, and adapt in the real world, not just generate text Have built production systems using LLMs, including experience with tool use, function calling, and multi-step orchestration Care deeply about reliability and safety when AI systems take real actions with real consequences Are comfortable across the stack—agent orchestration, backend services, and frontend interfaces Have strong Python and TypeScript skills (our primary languages) and can pick up new environments quickly Bring a product mindset: you think about what users need, not just what's technically interesting Communicate clearly and work well with a distributed team Strong candidates may also have: Experience with cloud infrastructure (AWS, Azure, GCP) or infrastructure-as-code tools Background in building developer tools, CLIs, or platform engineering systems Familiarity with container orchestration (Kubernetes, Docker) Experience building evaluation frameworks and benchmarks for AI systems Knowledge of prompt engineering, fine-tuning, or retrieval-augmented generation Compensation Base Salary: $190,000-220,000 All full-time employee offers at Pulumi include base salary, bonus or variable, equity and benefits (details below). Benefits Perks at Pulumi: Healthcare Coverage: We offer comprehensive medical, dental, vision, and supplemental insurance at no cost to U.S. employees. Internationally, we comply with local healthcare requ ... (truncated, view full listing at source)