Software and AI Engineer

Ford Pro is seeking a Software and AI Engineer to enhance Ford Motor Credit Company's mission-critical loan originations platform. The role involves diagnosing and improving multi-tier backend systems, building MCP tools and multi-agent workflows, developing knowledge graph and retrieval capabilities, and ensuring AI systems are auditable and production-ready.

Responsibilities:

• Read the whole stack, then change it. Diagnose production issues across the mainframe (COBOL, DB2, CICS), the Java/Spring services tier, and Cloud Run microservices. Ship the fix on the money-path and the corresponding improvement to agent tooling and knowledge substrate in the same change set
• Design and operate the MCP tools layer. Build least-privilege, per-call-audited MCP servers over DB2, Splunk, and GCP. Provenance, PII handling
• Build production multi-agent workflows on Google ADK — SequentialAgent and LoopAgent compositions, creating and managing Agent workflows, including escalation-containment patterns — for Splunk → DB2 → GCP error triage, ADR authoring, and originations debugging
• Own the knowledge graph and retrieval work end-to-end. Model the loan originations ontology (VIN, dealer, offer, program eligibility, environment topology). Implement hybrid vector + graph retrieval. Every RAG-produced answer must be defensible under audit
• Ship deterministic backend services in Java/Kotlin + Spring Boot on GCP — the audit envelope, provenance store, and human-in-the-loop review surfaces where non-determinism is unacceptable
• Author ADRs in ADR-driven-development, OKF business-logic documents, skill files, and Splunk saved-search catalogs. Pair with teammates on the hard debugging sessions. The write-up is part of "done."
• Track mean time from "agent cannot handle this" to "agent handles this autonomously," and publish the trend monthly
• Create evaluations and an evaluation runner for all AI tools and agents built by you and the team
• Publish, monthly, the change types you've made cheaper — mean time from "agent doesn't know this" to "agent handles it autonomously."
• Automate test-case generation, evaluation, and grading for both brownfield and greenfield applications, ensuring coverage, repeatability, and measurable quality improvements

Requirements:

• Bachelor's or Master's degree in Computer Science, Information Systems, or a related field — or equivalent professional experience
• 7+ years building and operating production-grade backend systems as a hands-on engineer
• 5+ years in Python and Java (Kotlin, TypeScript a plus)
• Atleast 1 year of experience in shipping AI-enabled applications in production — agent frameworks, retrieval systems, or LLM-backed services
• Demonstrated experience with agent orchestration frameworks — Google ADK, LangGraph, LangChain, or comparable
• Practical experience implementing MCP (Model Context Protocol) tool integrations, agent communication patterns, and AI observability
• Working experience with vector search and hybrid retrieval, and with knowledge graphs or GraphRAG
• Strong backend fundamentals: Spring Boot, relational databases (PostgreSQL, DB2, or similar) including query and schema design, REST/OpenAPI, TDD, modern CI/CD
• Production experience on Google Cloud Platform (Vertex AI, Cloud Run, BigQuery) or an equivalent cloud
• Comfort operating in complex, multi-tier, legacy-integrated environments where change control and auditability matter
• Candidates for positions with Ford Motor Company must be legally authorized to work in the United States
• Visa sponsorship is not available for this position
• Semantic technologies, ontology engineering, or graph databases (Neo4j, Neptune, or similar)
• Experience integrating with mainframe systems (COBOL, DB2 for z/OS, CICS) or comparable long-lived transactional cores
• Splunk SPL proficiency, including saved-search and dashboard authoring on large indexes
• Event-driven architectures (Pub/Sub, Kafka); containerization (Docker, Kubernetes)