Principal Agentic Engineer, AI Ecosystem

This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a Principal Agentic Engineer, AI Ecosystem based in the United States. The Principal Agentic Engineer will define and build the technical foundations of an organization-wide AI ecosystem, connecting AI platforms, developer tooling, and agentic product development. This Principal-level role combines strategic architecture ownership with hands-on engineering, ensuring that shared capabilities are designed consistently rather than developed independently across teams. You will establish reusable agent standards, execution frameworks, identity and authorization models, and evaluation infrastructure that enable safe, reliable, and scalable agent deployment. Working across multiple engineering groups and Product leadership, you will take ownership of critical technical challenges that fall between existing team responsibilities. You will also shape the long-term vision for an AI-native developer ecosystem while building and shipping foundational components yourself. This is an opportunity to influence technical direction at scale through architecture, prototypes, production code, and engineering mentorship. Accountabilities • Architect across the AI ecosystem: Maintain a system-wide architectural perspective across AI Platforms, AI Developer Experience (AI DevX), and Agentic Product Development Lifecycle (PDLC), ensuring that investments and technical capabilities work together effectively. • Define the technical vision and architecture for an AI-native developer ecosystem, enabling teams to build on shared primitives rather than independently solving the same problems. • Lead build-versus-buy decisions for PDLC-critical agents and ecosystem-wide capabilities, evaluating whether external solutions can deliver greater value than internal development. • Design and build foundational ecosystem components, including agent execution frameworks, context and retrieval systems, and telemetry that treats token consumption and cost as core design considerations. • Own cross-team foundational capabilities: Define what constitutes a reusable agent, including its unit structure, interfaces, and required standards. • Establish agent standards covering human-in-the-loop requirements, blast-radius policies, evaluation expectations, and promotion gates, partnering with AI Platforms for long-term infrastructure operations. • Design the non-human identity and agent authorization (authZ) model required to support scheduled and unattended agents across teams. • Build shared AI evaluation and regression infrastructure, including frameworks, CI integrations, and result storage, avoiding fragmented or duplicated implementations. • Deliver hands-on engineering: Write and ship production code for high-impact foundational components rather than limiting involvement to architecture recommendations or technical specifications. • Work closely with active initiatives, such as on-call investigation agents, to validate architectural decisions under real-world usage and economic constraints. • Prototype and de-risk technically challenging components before requesting engineering investment from other teams. • Lead through technical influence: Mentor Agentic Engineers and Staff Engineers across AI Platforms, AI DevX, and Agentic PDLC in agent architecture, context engineering, and evaluation practices. • Help strengthen the organization's senior technical capabilities through knowledge sharing and engineering guidance. • Produce architecture decision records, technical standards, and implementation playbooks that allow individual teams and their agents to build independently against a consistent architecture. • Represent a clear technical perspective in roadmap and build-versus-buy discussions, influencing decisions across teams without relying on formal management authority. • Partner closely with AI Platforms, AI DevX, Agentic PDLC, and Product leadership as a technical counterpart, with ownership focused on the shared challenges between these groups. Requirements • 12+ years of software engineering experience, including substantial work in platform or systems architecture. • Experience designing and delivering complex systems at consumer scale, preferably supporting 50 million or more monthly active users or an equivalent level of scale. • Demonstrated experience architecting and deploying production agentic systems, including designing the frameworks, identity models, and evaluation infrastructure on which other agents operate. • Direct experience with non-human identity, service-account management, or agent authorization models, or a well-reasoned technical approach to designing and implementing them. • Experience designing and consuming structured interfaces for AI systems, including MCP-style tool definitions, JSON Schema, and OpenAPI. • Understanding of tool-calling architectures and contract-first approaches to agent development. • Proven ability to build evaluation and regression frameworks for AI systems, including promotion gates, scoring mechanisms, and CI integration—not simply running ad hoc evaluations. • Strong architectural judgment across multiple services and engineering teams, with the ability to distinguish problems that require shared platform investment from those better addressed independently. • Strong understanding of AI economics, including token costs as an architectural constraint, progressive retrieval strategies, and the trade-offs between centralized and distributed model usage. • Demonstrated ability to influence technical direction across teams without direct reporting authority, using documentation, prototypes, and measurable engineering outcomes. • Excellent written communication skills, including the ability to produce precise architecture decision records, technical standards, and playbooks that other engineers and AI agents can follow