AI Engineer III

AI Engineer

This is a hybrid position in Plano, TX (first preference), Memphis, TN, or Pittsburgh, PA. Candidates residing within 50 miles of a Fed Ex campus will be required to work on-site at a Fed Ex location several times per week.

The AI Engineer is responsible for designing, developing, deploying, and maintaining artificial intelligence and machine learning solutions that support intelligent automation, predictive insight, and advanced analytics across the enterprise. As a hands-on builder, this role applies software engineering principles to write production-quality code, build scalable AI systems, including AI Agents and data pipelines, and integrate AI models into new and existing business applications.

The AI Engineer collaborates closely with Data Scientists, Data Engineers, ML Ops Engineers, and Platform teams to bring machine learning models from prototype to production. A critical part of this function is to ensure that AI use cases are transitioned from experimentation into reliable, governed, and business-ready solutions by owning their complete operational readiness. This includes implementing robust observability, defining Service Level Objectives (SLOs), and establishing clear incident response and rollback strategies for all AI services.

• Architect, develop, and maintain multi-agent systems, deterministic workflow agents, and autonomous tool-calling pipelines using the Google Agent Development Kit (ADK) in Python.
• Use the Google Agent CLI (agents-cli / adk) to scaffold, run, test, evaluate, and package agent projects across development, staging, and production environments.
• Implement complex agentic patterns including graph-based workflow runtimes, hierarchical multi-agent delegation, human-in-the-loop (HITL) tool confirmations, and state/session memory management.
• Build custom agent tools, Model Context Protocol (MCP) tool integrations, and API connectors powered by foundation models on Vertex AI.

• Design and implement high-performance RAG (Retrieval-Augmented Generation) architectures integrated with ADK using Vertex AI Vector Search, Big Query Vector Search, and Cloud Storage (GCS).
• Build scalable ingestion, chunking, and embedding pipelines using Python, Cloud Dataflow (Apache Beam), and Cloud Run Jobs.
• Integrate structured enterprise databases (Big Query, Cloud SQL / AlloyDB pgvector, Firestore) as tool-accessible data stores for autonomous agents.

• Deploy and operationalize ADK agents seamlessly to Vertex AI Agent Engine (Reasoning Engines), Cloud Run, and Google Kubernetes Engine (GKE) using the Agent CLI and containerized Python runtimes.
• Establish automated evaluation suites using agents-cli eval, Vertex AI GenAI Evaluation Service, and continuous regression testing for reasoning, tool fidelity, and context relevance.
• Implement production-grade agent observability, session telemetry, step-by-step tracing, and latency monitoring using Cloud Trace, Cloud Logging, and Vertex AI Model Monitoring.
• Define and own Service Level Objectives (SLOs) (e.g., TTFT, p50/p95 execution latency, tool invocation error rates, agent task completion rates).

• Secure agent tool execution, API calls, and data mutations using least-privilege IAM, VPC Service Controls, Secret Manager, and zero-trust safety guardrails.
• Expose agent capabilities via asynchronous FastAPI / gRPC microservices and register deployed agents into enterprise portals.
• Track rapid advancements in the Google agent ecosystem, including new ADK features, multimodal agents, and reasoning paradigms.

• Python Mastery:
  Expert proficiency in modern Python, asynchronous programming (asyncio), typing, structured data modeling, and web framework development (FastAPI).
• Google Agent Development Kit (ADK):
  In-depth experience with the google-adk Python SDK, including Task APIs, graph workflow runtimes, multi-agent coordination (sequential, parallel, loop, hierarchical), and session/memory layers.
• Google…