Forward Deployed Engineer

About the Role: As a Forward Deployed Engineer (FDE) at PubMatic, you will sit at the intersection of software engineering and applied AI, embedded directly within GTM and operational teams to ship production-grade AI workflows. You will be delivering working agents, integrations, and automation that measurably improve productivity and decision-making inside real business processes. Operating in a small pod model, you will own the full technical lifecycle: discovery, integration design, build, and production rollout. Initial focus areas include GTM functions where AI can automate analysis, generate insights, and streamline day-to-day operations. This is a hybrid role based in New York, with proximity to PubMatic’s GTM teams. On-site collaboration is expected for this role. Travel to Redwood City and other PubMatic locations may be required. Your goal will be to automate entire job workflows, not just small tasks. For example, you would enable teams to 10X the amount of workload they can handle (such as 10X the number of deals managed per person, or 10X the number of DSP integrations per person) by building end-to-end agents that manage entire workflows or processes today allowing people to focus on customer interaction and exception handling. The definition of success for this role is workload ÷ team size. What You'll Do: Technical Delivery & Production Deployment Own end-to-end technical delivery across multiple AI workflow deployments, from first prototype to stable production rollout. Understand and follow PubMatic Agentic development guidelines, Guardrails, Evals, to Build and deploy AI agents, workflow automation, and integration pipelines directly into operational systems used by PubMatic business teams. Develop MCP servers, sub-agents, retrieval pipelines, structured extraction templates, and supporting infrastructure tailored to each use case. Resolve data readiness, permissions, and system integration blockers to ensure solutions are production-ready rather than experimental. Ensure all deployed solutions meet reliability, latency, and observability requirements appropriate for enterprise operations. Use-Case Discovery & Embedded Collaboration Embed closely with functional leaders across GTM teams to identify and scope high-impact AI opportunities. Define success criteria and measurable outcomes for each workflow in collaboration with business stakeholders. Translate business requirements into technical designs, sequencing delivery to protect timelines and manage scope trade-offs. Guide functional teams through AI adoption, building trust, reducing friction, and ensuring deployed solutions are actively used. Platform & Repeatability Contributions Codify deployment patterns, connector templates, evaluation frameworks, and observability standards into reusable “golden path” playbooks. Feed field learnings back to PubMatic’s internal AI platform teams to accelerate the broader AgenticOS buildout. Contribute to evaluation design, precision/recall benchmarks, human review queues, audit trails, for each deployed workflow. Maintain strong working knowledge of LLM capabilities, agent development frameworks, and the evolving enterprise AI tooling landscape. Who You Are: We are looking for engineers who think in terms of workflow multipliers, not feature additions . Your mandate is to automate entire job workflows, enabling teams to handle 10X the volume of deals, integrations, launches, and tickets without adding headcount. The definition of success for this role is workload ÷ team size. Required 2+ years of software engineering or technical deployment experience that includes direct customer-facing or embedded delivery work. Hands-on experience with modern AI tooling including Claude, ChatGPT API, Notion AI, or comparable agent-development frameworks. Production experience with LLMs: prompt engineering, agent development, retrieval-augmented generation, evaluation frameworks, and deployment at scale. Strong programming profi…