The AI CoE builds AI products and the shared platform that powers them. We're looking for a Senior ML / AI Engineer who is equally comfortable building models and building the platform around them: someone
who can train and evaluate an ML model, ship a production LLM/GenAI application, and extend shared AI infrastructure and tooling that the wider team depends on.
This is a senior builder role at the intersection of applied data science, GenAI application engineering, and AI platform engineering. Our work spans predictive modeling, LLM-based systems, and the platform underneath them; you'll take problems from data and prototype through to governed, monitored production services, and set the technical patterns other engineers build on. Projects vary over time — we value engineers who can move across the stack rather than stay in one lane
Performance expectations:
- Build ML models — frame problems, engineer features, train and evaluate models (ranking, scoring, survival/time-to-event, classification, forecasting), and reason rigorously about metrics (AUC, C-index, calibration), validation strategy, subgroup performance, and failure modes.
- Integrate models into decision systems — combine model output with business/domain rules and LLM reasoning to produce explainable, trustworthy recommendations.
- Ship GenAI applications — design and deploy LLM-powered features: RAG pipelines, agents, structured extraction, summarization, decision-reasoning trails, and evaluation harnesses using Claude/Bedrock and other models.
- Engineer the AI platform — extend the shared AI gateway (unified multi-model access, API keys, per-team budgets, failover, observability) and reusable libraries/SDKs that other teams build on.
- Own the RAG/data layer — embeddings, vector stores, retrieval quality, chunking, and grounding strategies; measure and improve retrieval and answer quality.
- Build evaluation & quality tooling — offline/online eval, LLM-as-judge, regression suites, statistical validation, and guardrails so model and prompt changes ship safely.
- Productionize — wrap models and pipelines as tested, observable services (Python, containers, AWS Lambda/SageMaker/EKS), with monitoring for quality, cost, latency, and drift.
- Lead technically — set patterns and standards, review designs and code, mentor engineers, and partner with data scientists, MLOps, clinical/domain experts, and product owners to move prototypes to production.
Required Qualifications
- 5+ years building and shipping ML / AI systems in production (not just notebooks/POCs), including technical leadership of non-trivial projects.
- Strong data science / ML fundamentals — feature engineering, model training and evaluation, metrics (AUC, C-index, calibration, precision/recall), gradient-boosted trees (XGBoost), and sound experimental methodology (validation strategy, subgroup analysis).
- Experience building ranking, scoring, or survival/time-to-event models , and integrating model output into a larger decision system.
- Hands-on GenAI / LLM engineering — RAG, prompt engineering, function/tool calling, embeddings and vector search, and LLM evaluation.
- Excellent Python — production-grade, tested, well-structured code; comfortable building APIs/services and shared libraries.
- AWS experience — Bedrock and/or SageMaker, Lambda, S3, plus containers (Docker) and Git-based workflows.
- Solid software engineering practice: version control, testing, code review, CI/CD; ability to reason about cost, latency, and reliability of AI systems in production.
Preferred Qualifications
- AI platform engineering — building shared gateways/proxies, model routing, multi-tenancy, quota/budget enforcement, or internal AI SDKs.
- Experience with agent frameworks, real-time/voice AI, or streaming inference.
- Vector databases (Qdrant, OpenSearch, pgvector) and retrieval-quality tuning at scale.
- IaC (Terraform), observability (OpenTelemetry/CloudWatch), and FinOps for AI workloads.
- Healthcare / clinical ML — survival analysis, outcome prediction, or working with clinical/scientific datasets alongside domain experts.
- Serving models as endpoints (SageMaker), cross-account inference, and train/serve parity.
- Experience in a regulated / PHI-handling environment (HIPAA) — data governance, PII handling, auditability.