About the Role
We are hiring a Principal Platform Engineer, AI to design, build, and operate the enterprise AI platform for Hasbro and Wizards of the Coast — the shared foundation that becomes the standard, governed path for how internal teams access and leverage AI capabilities.
What You'll Build
- Enterprise AI runtime: Model gateway, routing layer, and provider abstraction with identity-aware access, quota controls, and end-to-end traceability.
- Agent and Tool Protocol Standards: MCP and A2A as the connective tissue between agents, tools, and data sources, including authentication/authorization and security patterns for each.
- Responsible AI: Identity-aware access controls, prompt management standards, evaluation frameworks, central policy controls, and safety guardrails that adjust without app-side code changes.
- Data and Retrieval Integration: Data platforms (Databricks, Snowflake, etc.) integration, grounding and context patterns that fully exploit modern model capabilities supercharged with a wealth of owned data.
- Enterprise AI Administration: Own configuration, governance, and operational management of enterprise AI systems, ensuring consistent policy, auditability, and cost visibility across managed AI products.
- Enterprise AI Operations: Cost attribution, rate limiting, audit telemetry, and operational monitoring designed to be operational from day one.
How You'll Succeed
This is not a greenfield position; you'll be joining an an organization where AI investment is already substantial. Structured teams are running production workloads, analytics and research organizations with deep Databricks and Snowflake investments, and yes, shadow deployments, rogue MCP servers, and bespoke integrations that teams built because there wasn't yet a useful official path. This demand and momentum, and the teams behind them, will be critical to partner with.
This role needs someone who can walk into this environment and become the clear center of gravity, not through authority but through credibility, good defaults, and a platform that is genuinely easier to use than what teams have already built. The technical work matters, but excelling in this role will depend on your ability to bring people with you.
You Will
- Partner across the organization to drive agreement on interface contracts, especially around MCP and A2A protocols, identity, and data access, so teams can build in parallel rather than waiting on each other.
- Earn adoption from research, engineering, and product teams already engaged with AI, by making your path obviously better, not by mandate. You'll meet teams where they are, respect what they've already shipped, and sequence migrations with emotional intelligence.
- Work with security, legal, and risk partners to translate AI safety from policy into running code that holds up in audit and in production.
- Drive consensus through written design docs, evaluation results, and decision records, not through status meetings or organizational pressure.
- Communicate at every altitude, from an on-call incident to a VP review — without losing precision or changing register.
What You'll Bring
- 10+ years in platform, infrastructure, or systems engineering with direct ownership of production systems used by other engineers.
- Hands-on experience building shared AI platforms, specifically with AWS Bedrock, AgentCore, and/or SageMaker, for multi-team use at scale.
- Deep familiarity with agent and tool protocols including MCP and A2A, including their authentication and security models.
- Working knowledge of enterprise identity and access control (Microsoft Entra ID, Okta) and how IAM patterns apply to AI systems and agent-to-agent communication.
- Experience integration with modern data platforms (Databricks, Snowflake, etc.) and building grounding and retrieval patterns using tool-use and function calling.
- Familiarity with enterprise AI administration platforms such as Claude for Enterprise and ChatGPT Enterprise, including governance, policy configuration, and cost management.
- Azure and/or Databricks experience is a meaningful plus, especially when in context of supporting analytics and research-heavy organizations.
- Pragmatic about AI safety; you treat guardrails as an engineering discipline rather than a compliance checkbox.
- A proven track record of driving cross-team technical adoption through credibility, clarity, and well-built defaults rather than relying on organizational authority.