Staff Data Engineer - Agentic AI, AI Orchestration & Data Platforms
General Motors
- Location
- Austin, Texas, United States | Warren, Michigan, United States
- Track
- AI Agent Engineer
- Level
- Staff
- Salary
- $160K–$211K / yr
- Posted
- November 19, 2026
- Source
- Built In
Job description
Description
This role is categorized as hybrid. This means the successful candidate is expected to report to Warren Global Technical Center or Austin Technical Center three times per week, at minimum [or other frequency dictated by the business if more than 3 days].
The Role
The Quality & Safety Data Platform team builds the governed data products behind GM quality, warranty, dealer, and vehicle analytics. We are hiring a hands-on Level 8 individual contributor to lead modernization of those platforms on Azure Databricks and to make agentic AI part of how the team builds, validates, and operates them. You will own the architecture and delivery of lakehouse pipelines, and you will design agents, reusable skills, and orchestration workflows that speed up legacy ETL migration, data validation, and production support. You will set technical direction across QDAP, CARD, MTDR, GART, Compass, and PRTS, mentor senior engineers, and influence engineering standards beyond this team. Data engineering is the job. Agentic AI is how this engineer multiplies the team's impact.
What You'll Do
- Architect and productionize reliable batch and streaming data products in Azure Databricks, using Spark Declarative Pipelines, Delta Lake, Unity Catalog, and Bronze, Silver, and Gold patterns that quality, warranty, and AI applications can trust.
- Lead migration of legacy ETL, including IBM DataStage, Oracle, DB2, and on-prem platforms, to incremental Databricks pipelines, with cutover plans, parity criteria, and automated reconciliation against the source system.
- Design, build, and operate AI agents and reusable skills that automate code conversion, pipeline generation, validation triage, troubleshooting, and documentation, and publish them so other teams can adopt them.
- Engineer agent orchestration for stateful multi-step and multi-agent workflows, including tool calling, memory and context, human approval, retries, and a deterministic path when an autonomous agent is the wrong tool.
- Prepare governed data for AI, including retrieval and vector search, evaluation datasets, lineage, and access controls, so agents and users work from trusted, current data.
- Set the engineering bar for CI/CD, workload identity, security, cost, observability, and production support, and gate agent releases with evaluation of task success, tool use, grounding, latency, and cost.
- Set technical direction across teams, mentor senior engineers, and partner with quality, warranty, aftersales, and platform teams to turn ambiguous needs into measured outcomes.
Your Skills & Abilities (Required Qualifications)
- Bachelor's degree in Computer Science, Software Engineering, Data Engineering, or a related field, or equivalent experience.
- 8+ years of professional experience in data engineering, software engineering, or platform engineering, including technical leadership of production enterprise data platforms.
- Expert proficiency in Python, SQL, and Apache Spark (PySpark), with production-quality software engineering practices (testing, code review, version control, modular design).
- Deep, current hands-on experience with Databricks: Delta Lake, Unity Catalog, Workflows/Jobs, Spark Declarative Pipelines (DLT), Asset Bundles, and medallion or comparable lakehouse architecture.
- Proven experience leading legacy ETL or data warehouse modernization to a cloud lakehouse, including data reconciliation and parity validation against the source system.
- Hands-on experience designing, building, and deploying AI agents or LLM-powered automation with tool calling and orchestration, using at least one agent framework or SDK (for example, Claude Agent SDK, OpenAI Agents SDK, LangGraph, LangChain, CrewAI, Databricks Mosaic AI Agent Framework) or equivalent custom code.
- Daily, hands-on use of AI-assisted development tools such as Cursor, Claude Code, or GitHub Copilot, including authoring agent instructions, rules, and reusable skills that improve engineering productivity.
- Hands-on experience with RAG or vector search, and with evaluating LLM or agent behavior using golden datasets, tool-use checks, or LLM-as-judge methods.
- Experience with CI/CD and GitHub-based deployment of data products, including GitHub Actions, environment promotion, and secure workload identity (OIDC/service principals).
- Experience with a major cloud platform (Azure preferred: ADLS Gen2, Key Vault, Entra ID; AWS or GCP also considered).
- Solid understanding of data modeling, CDC/incremental processing, streaming, performance tuning (partitioning, clustering, file layout), data quality, governance, and lineage.
- Demonstrated ability to work independently through ambiguity, set direction across teams, influence without direct authority, and deliver measurable improvements in quality, reliability, cost, or business value.
- Strong written and verbal communication skills, including technical documentation and executive-level summaries.
What Can Give You a Competitive Advantage (Preferred Qualifications)
- Experience with the Model Context Protocol (MCP), including building MCP servers or connectors, and with multi-agent patterns such as planner-executor, supervisor-worker, or agent-to-agent handoff.
- Experience building agent evaluation harnesses and observability (for example, MLflow Tracing, LangSmith, Langfuse, Promptfoo, Ragas, or DeepEval) and wiring them into CI/CD gates, including drift checks on production traces.
- Experience with agent guardrails: prompt-injection defense, scoped credentials, input and output validation, PII controls, sandboxed execution, human approval, and an audit trail of agent decisions.
- Experience routing across model tiers to balance quality, latency, and token cost, and with short-term and long-term agent memory.
- Experience preparing unstructured data for retrieval: chunking, embeddings, hybrid search, reranking, and retrieval evaluation.
- Experience with Databricks Genie, Mosaic AI Vector Search, Model Serving, AI Gateway, D