Job DescriptionOur Benefits: Our team members enjoy a robust benefits package including:
- Competitive compensation
- Paid time off for vacations, holidays, and illness
- Comprehensive health insurance (medical, dental, vision, prescription)
- 401(k) retirement savings plan
- Education support and student loan assistance
- Life and disability insurance
- Flexible spending account
Job SummaryThe AI Engineer Specialist serves as a senior hands-on Subject Matter Expert (SME) within the CHS Enterprise AI Center of Excellence (CoE), operating under a collaborative hub-and-spoke model. Bridging deep machine learning engineering with modern agentic architecture, this role builds reusable "golden-path" blueprints, customizes foundation and domain-specific models on Google Cloud (Vertex AI), and embeds automated safety and observability guardrails into production. You will solve complex, high-stakes healthcare and clinical AI challenges across Generative AI, Agentic workflows, and Predictive/ML patterns while enabling business and technology spoke teams to scale safely across the enterprise.
Key Responsibilities- Reference Architectures & "Golden Path" Engineering: Design, build, and harden reusable, pre-secured code templates and component libraries across CHS's three core AI patterns: GenAI (grounded assistants), Agentic workflows, and Predictive/ML pipelines.
- Model Customization & Clinical NLP Optimization: Evaluate when foundation models (via Vertex AI Model Garden, including Gemini and open-weight models) require domain adaptation. Lead parameter-efficient fine-tuning (PEFT/LoRA), retrieval/grounding optimization, and clinical terminology integration (SNOMED-CT, RxNorm, ICD-10/11) for high-precision healthcare use cases.
- Agentic Systems & AI-for-BI Delivery (Spoke Execution): Serve as a lead hands-on builder for CoE-led flagship initiatives-engineering multi-agent orchestration (using Google ADK, Agent Runtime, Sessions/Memory Bank), Model Context Protocol (MCP) tool servers on Cloud Run, and enterprise AI/BI data agents (e.g., Looker Data Agents and master orchestrators).
- Guardrails as Code & Automated Evaluation: Operationalize AI Governance Committee (AIGC) policies into automated pipeline controls-implementing Model Armor ingress/egress scanning, sensitive-data (PHI/PII) filters, custom Human-in-the-Loop (HITL) approval workflows, and automated quality/groundedness/drift evaluation suites.
- AI Economics, Performance & Observability: Optimize latency, throughput, and token efficiency across multi-step agent runs. Instrument all solutions with OpenTelemetry traces, HIPAA/CMS compliance audit logs (capturing prompts, tool calls, and data access), and FinOps cost-per-solution telemetry routed through the Enterprise AI Gateway.
- Spoke Enablement & Readiness Stewardship: Partner with business and technology spoke teams through design reviews, developer office hours, and structured production-readiness milestones (Design Review, Non-Prod Readiness, Production Promotion, and Post-Deployment Review) to accelerate safe launches.
Minimum Qualifications- Education: Bachelor's degree in Computer Science, Data Science, Machine Learning, Biomedical Informatics, or a related quantitative engineering field required; Master's degree preferred.
- Experience: 8-10 years of deep Machine Learning, NLP, and AI engineering experience, including hands-on experience designing and shipping production Generative AI, RAG, or multi-agent applications into enterprise cloud environments.
Knowledge, Skills and Abilities- Strong proficiency in Python, PyTorch, and the Hugging Face ecosystem, paired with hands-on mastery of GCP Vertex AI (Model Garden, Agent Development Kit [ADK], Agent Gateway, Model Armor, Cloud Run) and working familiarity with Azure/Copilot environments.
- Deep experience engineering production CI/CD pipelines for AI workloads with strict dev/QA/prod environment separation, least-privilege IAM/agent identity controls, and OpenTelemetry-based observability and token cost tracking.
- Working knowledge of healthcare IT standards and clinical terminologies (FHIR, HL7, SNOMED-CT, RxNorm, ICD-10/11), combined with practical experience enforcing HIPAA and CMS data privacy and auditability requirements.
- Ability to translate abstract clinical or business use cases into well-architected multi-step agent workflows, balancing model accuracy, latency, and unit economics to deliver reliable, production-ready solutions.
This is a fully remote opportunityThis position is not eligible for immigration sponsorship now or in the future. Applicants must be authorized to work in the U.S. for any employer.