Software Engineer III, Applied Agentic AI

About Us:   Proofpoint is a global leader in human- and agent-centric cybersecurity. We protect how people, data, and AI agents connect across email, cloud, and collaboration tools. Over 80 of the Fortune 100, 10,000 large enterprises, and millions of smaller organizations trust Proofpoint to stop threats, prevent data loss, and build resilience across their people and AI workflows. Our mission is simple: safeguard the digital world and empower people to work securely and confidently. Join us in our pursuit to defend data and protect people. How We Work: At Proofpoint you’ll be part of a global team that breaks barriers to redefine cybersecurity guided by our BRAVE core values:  Bold in how we dream and innovate Responsive to feedback, challenges and opportunities Accountable for results and best in class outcomes Visionary in future focused problem-solving Exceptional in execution and impact The Role   Proofpoint is hiring an Agentic AI Engineer on the team building our agentic AI platform — the agent runtimes, MCP access layer, and evaluation tooling that product teams across the company build on. You'll be building for other engineers: the people shipping agentic features into products that thousands of enterprises rely on around the clock.   This role sits at the frontier of agentic technology. New models, frameworks and protocols land faster than any single product team can absorb, and part of your job is to absorb them first: work out what actually holds up, establish the best practices, and prove them out in reference agentic applications that other teams can learn from and build on. You get first look at the new stuff, and the fun part of the job is showing engineers across the company what it can do.   What you’ll do   Build reference agentic applications on our platform — end-to-end examples of how agents should be designed, evaluated, deployed and monitored at Proofpoint.  Establish and document best practices: tool design and MCP integration, prompt and context management, memory and planning, cost/latency tradeoffs, failure recovery, and what "release-ready" means for a nondeterministic system.  Design evaluation harnesses — task suites, baselines, and quality, cost and latency metrics — and use them to decide what ships.  Turn ambiguous questions into working prototypes quickly, and present results to the teams who asked; recommend just as readily when an approach should be abandoned.  Harden successful prototypes into reusable platform capabilities that product teams can adopt.  Partner with experienced engineers, security researchers and product tech leads to go deep on the systems you build on.  Socialize agentic patterns across the company through demos, design reviews, written guides, and pairing sessions.  Communicate and understand stakeholders' requirements and participate in cross-team design discussions.   What we’re looking for   Bachelors or Master's Degree in Computer Scien…