Principal Software Engineer

About the role

Forward Deployed Engineers are embedded directly with Google’s enterprise customers and serve as their primary technical point of contact across the full delivery lifecycle. FDEs own the relationship as much as the technology — translating business needs into AI solutions, anticipating blockers before they surface, and communicating progress with clarity to both executive sponsors and engineering teams. This is a customer-facing role where technical excellence and outstanding communication are equally non-negotiable.

Delivery lifecycle ownership

• Engage: Requirements & discovery — lead stakeholder workshops, ask the right questions proactively, and document requirements before the customer thinks to ask. Establish trust and communication cadence from day one.
• Design: User journey design — translate business needs into conversation flows, escalation paths, and agent skill boundaries — presenting options clearly to non-technical sponsors and technical teams alike.
• Build: Agent development — implement GECX agents end-to-end: prompt engineering, RAG pipelines, API integrations, GCP infrastructure. Proactively flag scope risks and communicate progress at every sprint.
• Validate: Testing & evaluation — build automated eval frameworks and define acceptance criteria with the customer. Communicate quality status transparently — no surprises at go-live.
• Ship: Deployment & go-live — lead production deployments with clear cutover comms to all stakeholders. Proactively anticipate go-live issues and brief executive sponsors before they ask.
• Support: Post go-live support — stay embedded through stabilization. Surface performance insights proactively, communicate incident status in plain language, and keep customer confidence high.
• Grow: Enablement & improvement — train and empower customer teams to own their GECX agents. Document patterns clearly.

Ideal candidate

• Exceptional communicator — translates architecture to execs and debug logs to engineers, same day.
• Proactive by default — surfaces risks, blockers, and insights before being asked.
• Customer-first mindset — builds trust through transparency, responsiveness, and follow-through.
• Full-lifecycle owner — taken AI deployments from discovery through hypercare, end-to-end.
• Agent-native builder — shipped agents running unsupervised in production.
• RAG practitioner — re-ranking, hybrid search, advanced context optimization.
• GCP fluent — Vertex AI, Cloud Run, BigQuery, IAM daily vocabulary.
• Quality-obsessed — defines done as: tested, evaluated, monitored, and handed over cleanly.