Job Description & Details
This is a hands-on senior AI engineering gig focused squarely on building enterprise-grade AI within the Microsoft ecosystem. If you enjoy wrangling Azure AI Foundry, setting up governance guardrails, and architecting multi-agent systems for regulated spaces, this role cuts straight through the noise.
What You'll Actually Be Doing
Expect to spend your days deploying and operating AI solutions, configuring evaluation pipelines, and setting up responsible AI controls. You'll be designing multi-step agentic workflows and tool-enabled agents, which means figuring out how autonomous orchestration patterns actually behave without breaking production. You'll also spend plenty of time collaborating with architects to map out data flows, secure identity integrations, and figure out why a particular prompt is blowing past token limits or hallucinating edge cases.
The Core Tech Stack
You need deep, practical fluency with Azure AI Foundry and the broader Microsoft Copilot ecosystem, including Copilot Studio and M365 integrations. On the coding side, you'll be leaning heavily on Python and TypeScript or JavaScript to stitch together APIs, data services, and custom agents. Don't apply if you've only tinkered with OpenAI wrappers in a notebook; they need someone who genuinely understands enterprise security, identity management, and how to govern AI in a locked-down cloud environment.
Interview Expectations
They're going to push you on system design for autonomous agents, likely asking how you handle failure states and state management when a multi-step agentic workflow stalls halfway through a complex task. The hiring manager wants to see if you can reason about cost, latency, and security trade-offs, not just whether you know which API endpoint to call. Expect a deep dive into how you implement responsible AI controls and evaluation pipelines in production without choking developer velocity.
Application Advice
Your resume needs to scream Azure and enterprise governance. Make sure keywords like Azure AI Foundry, agentic workflows, prompt engineering, Python, and TypeScript are front and center. Skip the generic buzzwords and instead highlight specific instances where you took an AI model from a rough proof-of-concept to a secure, compliant deployment in a regulated cloud environment.