Senior Generative AI Engineer

Light & Wonder • Austin, TX

2d • Onsite

About The Position

We are building a tight-knit, senior engineering group based in Austin, TX, tasked with creating the next generation of enterprise automation at Light & Wonder. Our mission is to design, deliver, and scale production-grade Agentic AI workflows that execute highly complex, meaningful tasks across the business. We will operate like a high-functioning startup within the enterprise: we favor shipping over process, rigorous evaluation over opinions, and strategic platform investments that make every subsequent deployment faster. We are leaning heavily into the modern Microsoft ecosystem, anchoring our architecture on the newly GA Microsoft Agent Framework (.NET), Azure OpenAI, and an advanced data backend powered by Snowflake, Databricks, and Microsoft Fabric. If you want to build resilient, multi-agent systems at enterprise scale, this is the team. We are looking for highly capable product builders who will own complex agentic workflows from the user interface down to the model execution. You will navigate across a React front-end, a C#/.NET orchestration layer, and various tool integrations to deliver agents that feel like refined, intelligent products. You will not be confined to a single layer of the stack. You will be responsible for the orchestration logic, multi-agent handoffs, and the rigorous evaluation coverage, ensuring your agents behave predictably in production. We don't expect you to have years of experience with the Microsoft Agent Framework specifically. We do expect broad technical expertise, a deep understanding of modern async patterns, and the ability to ship resilient AI software.

Requirements

• 6+ years of professional software engineering experience, with meaningful production depth on both front and back ends
• Strong proficiency in C# and .NET for server-side work, including modern .NET (8 / 9 / 10), async patterns, and dependency injection. This is the primary stack and is required, not preferred
• Strong proficiency in TypeScript and React for front-end work
• Experience shipping at least one production system that uses LLMs end-to-end. "Production" means real users, real reliability constraints, and real evals - not a hackathon project or a tutorial follow-along.
• Multi-step orchestration or agent experience preferred over single-completion-call experience
• Demonstrated ability to learn unfamiliar frameworks quickly and write production-grade code while doing so. We will probe for this in the technical screen
• Strong grasp of API design, auth, async patterns, and the tradeoffs of streaming versus request-response
• Working knowledge of Azure primitives (App Service / AKS / Functions, Key Vault, API Management, Entra ID) and comfort writing IaC (Bicep)
• Experience writing evals for non-deterministic systems: behavioral tests, regression harnesses, human-rated samples

Nice To Haves

• Hands-on time with Microsoft Agent Framework, Semantic Kernel, or AutoGen. MAF specifically is recently GA; depth is uncommon and not required
• Experience with .NET Aspire and Azure Functions hosting patterns for agents
• Experience with AG-UI, CopilotKit, or comparable agent/copilot UI frameworks
• Experience with N8N, Make, Zapier, or similar low-code / automation platforms at production scale
• Experience with Foundry Observability and Foundry Evals, including the Azure DevOps Foundry-evals extension for CI/CD
• Prior experience on a platform or developer experience team. You know what reusable looks like
• Exposure to enterprise identity (Entra ID / AAD), data classification, and DLP patterns
• Familiarity with MCP (Model Context Protocol) and A2A (Agent-to-Agent) interoperability patterns

Responsibilities

• Architect and deliver production-grade agent workflows end-to-end, spanning conversational UI, orchestration logic, and tool integrations
• Build intuitive, embedded agent interfaces using AG-UI and CopilotKit to deliver a refined product experience
• Implement robust orchestration using Microsoft Agent Framework (.NET), handling tool calling, multi-agent patterns, and human-in-the-loop handoffs
• Integrate with low-code platforms like N8N for workflow edge cases to accelerate delivery without sacrificing reliability
• Develop comprehensive behavioral, regression, and adversarial evaluations for all shipped workflows to guarantee production reliability
• Collaborate with Data Engineering on consumable data surfaces and with Platform Engineering on deployment and observability standards
• Contribute to shared platform primitives, reusable templates, and rigorous code review practices across the team

Benefits

• The company is committed to the highest standards of integrity, from promoting player responsibility to implementing sustainable practices.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: Associate degree