Staff Machine Learning Engineer, Platform (MLOPS)

At Credit Acceptance, we are building on the foundation of a successful business while investing in the future through technology, data, AI, and continuous improvement. As we continue to evolve, our Team Members play an active role in improving experiences, driving meaningful change, and shaping the future of our business. We foster an environment where people are empowered to contribute ideas, develop their careers, make an impact, and help lead our continued transformation.

Our Engineering teams are helping shape the future of Credit Acceptance through innovation. Team Members collaborate across the organization to solve complex challenges, build impactful solutions, challenge the status quo, and support informed business decisions. We value curiosity, continuous learning, and new ideas from people who are passionate about driving modernization and meaningful business impact.

Outcomes and Activities:

• •Own the deployment path for ML and GenAI models end to end training and inference pipelines, model registry and versioning, serving endpoints, and controlled promotion across development, QA and production.
• •Own runtime health for models in production: monitoring, alerting, drift and quality-regression detection, latency and throughput objectives, capacity and autoscaling behavior, and incident response through to root cause and a closed corrective action.
• •Operate the agent runtime layer. Route production agents through the enterprise AI Gateway and MCP Gateway rather than direct model and tool access, migrate existing agents onto that governed path, and keep tool surfaces scoped, versioned and least-privilege as they change.
• •Own evaluation of agents in production, not just before release. Online scoring and behavioral monitoring, quality-regression and drift detection against pinned baselines, sampling and judge pipelines, and the release-gate mechanics that stop a regression from shipping.
• •Build and maintain the observability and evaluation substrate other teams depend on: trace and telemetry capture including multi-turn and multi-step agent traces, logging standards, evaluation pipeline plumbing, and the data contracts underneath them.
• •Own platform unit economics. Measure and manage cost per inference, per document and per interaction, and produce the platform and infrastructure cost analysis that informs build-versus-buy and hosting decisions.
• •Make the paved road real. Deliver reusable pipeline templates, deployment patterns, reference implementations and internal tooling so product teams adopt the standard path because it is faster, not because it is mandated.
• •Partner with Cloud Engineering, Data Engineering, Security and SRE so the ML platform sits inside enterprise governance, identity and observability rather than beside it.
• •Respond to AI-specific production incidents and drive them to a closed corrective action: prompt injection attempts, rogue-agent cost spikes, data-classification exposure through a tool call, delegation abuse between agents, and model endpoint failures.
• •Maintain the architecture documentation and system diagrams for the ML platform and keep them accurate enough to be used in design review.
• •Mentor engineers and interns on production ML practice and raise the operating standard through design and code review rather than through rework.

Competencies: The following items detail how you will be successful in this role.

• •Customer Empathy: Customer Empathy is the ability to understand the perspectives, pain points, and experiences of customers. It involves actively putting oneself in the customer's shoes, comprehending their needs and challenges, and using that understanding to provide a better, more customer-centric experience.
• •Engineering Excellence: Engineering Excellence is about bringing great craftsmanship and thought leadership to deliver an outstanding product that delights customers and solves for the business. This involves the pursuit and achievement of high standards, best practices, innovation, and superior solutions.
• •One Team: A One Team mindset refers to a collaborative approach across the organization, where individuals work together seamlessly, without boundaries, as a single, cohesive team. Shared goals, open communication and mutual support create a sense of collective purpose. This enables teams to navigate challenges and pursue shared objectives more effectively.
• •Owner's Mindset: Owner's Mindset involves adopting a set of behaviors that reflect a sense of responsibility, accountability, strategic thinking, and a proactive approach to managing your domain. As an owner, you understand the business and your domain(s) deeply and solve for the right outcome for the domain(s) and the business.

Requirements

• •Bachelor's degree in Computer Science, Engineering, Statistics or a relevant technical field with at least 7 years of relevant experience, or a Master's degree in one of those fields with at least 5 years of relevant experience.
• •5+ years building and operating production ML or AI systems, with direct ownership of at least two of: training or inference pipelines, model serving infrastructure, model registry and versioning, or production monitoring and alerting.
• •Demonstrated ownership of a production ML or AI service through its full operating life: deployed it, monitored it, was paged for it, diagnosed a failure, and shipped the fix.
• •Strong Python and SQL, with production-quality engineering practice: version control, testing, code review, and CI/CD applied to ML workloads rather than only to application code.
• •Hands-on experience with a cloud ML platform in production. AWS and Databricks strongly preferred, including model serving, job orchestration, and a model registry or experiment tracking system such as MLflow.
• •Working knowledge of how LLM and GenAI workloads differ operationally from traditional ML: token cost and latency behavior, caching and batching, non-deterministic output.
• •Experience running LLM or agent applications behind a gateway or proxy layer, and able to build best practices for model routing and fallback, credential and key management, rate limiting and budget enforcement.
• •Working knowledge of tool-calling architecture for agents, including the Model Context Protocol: what an MCP server is, how tools are scoped and authorized, and why tool access is brokered through a gateway rather than granted directly.
• •Experience with containerization and infrastructure as code.
• •Ability to communicate technical and non-technical trade-offs clearly in writing to an audience that includes both engineers and non-engineers.
• •Candidates must reside in or be willing to relocate to one of the following hub locations: the greater Southfield, Michigan, area; the greater Bellevue, Washington, area; or the San Francisco Bay Area, California.

Preferred

• •Production experience with GPU-backed model serving, including autoscaling behavior under concurrent load, cold-start management and cost control.
• •Experience with OpenTelemetry and an enterprise observability platform such as Dynatrace, including instrumenting AI workloads rather than only conventional services.
• •Experience with model serving efficiency techniques: quantization, parameter-efficient fine-tuning, distillation or inference optimization.
• •Experience with data governance and access control on a lakehouse platform, such as Databricks Unity Catalog.
• •Experience operating AI systems in a regulated industry, particularly financial services, including auditability, retention and access-control requirements.
• •Experience with agentic or multi-step AI systems in production, including tracing and debugging multi-turn behavior.
• •Hands-on experience with a managed agent or tool gateway, such as AWS Bedrock AgentCore Gateway, and with running MCP servers in a governed environment.
• •Familiarity with emerging agent interoperability and agent identity standards, including Agent2Agent (A2A) style agent-t