Principal Data Engineer – ML Platforms

Altarum | Data & AI Center of Excellence (CoE) Altarum is building the future of data and AI infrastructure for public health - and we’re looking for a Principal Data Engineer – ML Platforms to help lead the way. In this cornerstone role, you will design, build, and operationalize the modern data and ML platform capabilities that power analytics, evaluation, AI modeling, and interoperability across all Altarum divisions. If you want to architect impactful systems, enable data science at scale, and help ensure public health and Medicaid programs operate with secure, explainable, and trustworthy AI - this role is for you.   What You'll Work On This role blends deep engineering with applied ML enablement:   ML Platform Engineering: modern lakehouse architecture, pipelines, MLOps lifecycle  Applied ML enablement: risk scoring, forecasting, Medicaid analytics  NLP/Generative AI support: RAG, vectorization, health communications Causal ML operationalization: evaluation modeling workflows Responsible/Trusted AI engineering: model cards, fairness, compliance    Your work ensures that Altarum’s public health and Medicaid programs run on secure, scalable, reusable, and explainable data and AI infrastructure.  What You'll Do

Platform Architecture & Delivery  

Design and operate modern, cloud-agnostic lakehouse architecture using object storage, SQL/ELT engines, and dbt. 

Build CI/CD pipelines for data, dbt, and model delivery (GitHub Actions, GitLab, Azure DevOps). 

Implement MLOps systems: MLflow (or equivalent), feature stores, model registry, drift detection, automated testing. 

Engineer solutions in AWS and AWS GovCloud today, with portability to Azure Gov or GCP. 

Use Infrastructure-as-Code (Terraform, CloudFormation, Bicep) to automate secure deployments. 

Build scalable ingestion and normalization pipelines for healthcare and public health datasets, including: 

FHIR R4 / US Core (strongly preferred) 

HL7 v2 (strongly preferred) 

Medicaid/Medicare claims & encounters (strongly preferred) 

SDOH & geospatial data (preferred) 

Survey, mixed-methods, and qualitative data 

Create reusable connectors, dbt packages, and data contracts for cross-division use. 

Publish clean, conformed, metrics-ready tables for Analytics Engineering and BI teams. 

Support Population Health in turning evaluation and statistical models into pipelines. 

Define SLOs and alerting; instrument lineage & metadata; ensure ≥95% of data tests pass. 

Perform performance and cost tuning (partitioning, storage tiers, autoscaling) with guardrails and dashboards. 

Build production-grade pipelines for risk prediction, forecasting, cost/utilization models, and burden estimation. 

Develop ML-ready feature engineering workflows and support time-series/outbreak detection models. 

Integrate ML assets into standardized deployment workflows. 

Build ingestion and vectorization pipelines for surveys, interviews, and unstructured text. 

Support RAG systems for synthesis, evaluation, and public health guidance. 

Enable Palladian Partners with secure, controlled-generation environments. 

Translate R/Stata/SAS evaluation code into reusable pipelines. 

Build templates for causal inference workflows (DID, AIPW, CEM, synthetic controls). 

Support operationalization of ARA’s applied research methods at scale. 

Implement Model Card Protocol (MCP) and fairness/explainability tooling (SHAP, LIME). 

Ensure compliance with HIPAA, 42 CFR Part 2, IRB/DUA constraints, and NIST AI RMF standards. 

Enforce privacy-by-design: tokenization, encryption, least-privilege IAM, and VPC isolation. 

Develop runbooks, architecture diagrams, repo templates, and accelerator code. 

Pair with data scientists, analysts, and SMEs to build organizational capability. 

Provide technical guidance for proposals and client engagements. 

Platform skeleton operational: repo templates, CI/CD, dbt project, MLflow registry, tests. 

Two pipelines in production (e.g., FHIR → analytics and claims normalization). 

One end-to-end CoE lighthouse MVP delivered (ingestion → model → metrics → BI). 

Completed playbooks for GovCloud deployment, identity/secrets, rollback, and cost control. 

Pipeline reliability meeting SLA/SLO targets. 

≥95% data tests passing across pipelines. 

MVP dataset onboarding ≤ 4 weeks. 

Reuse of platform assets across ≥3 divisions. 

Cost optimization and budget adherence. 

What You'll Bring

7–10+ years in data engineering, ML platform engineering, or cloud data architecture. 

Expert in Python, SQL, dbt, and orchestration tools (Airflow, Glue, Step Functions). 

Deep experience with AWS + AWS GovCloud. 

CI/CD and IaC experience (Terraform, CloudFormation). 

Familiarity with MLOps tools (MLflow, Sagemaker, Azure ML, Vertex AI). 

Ability to operate in regulated environments (HIPAA, 42 CFR Part 2, IRB). 

Experience with FHIR, HL7, Medicaid/Medicare claims, and/or SDOH datasets. 

Databricks, Snowflake, Redshift, Synapse. 

Event streaming (Kafka, Kinesis, Event Hubs). 

Feature store experience. 

Observability tooling (Grafana, Prometheus, OpenTelemetry). 

Experience optimizing BI datasets for Power BI. 

Logistical Requirements

At this time, we will only accept candidates who are presently eligible to work in the United States and will not require sponsorship.

Our organization requires that all work, for the duration of your employment, must be completed in the continental U.S. unless required by contract.

If you’re near one of our offices (Arlington, VA; Silver Spring, MD; or Novi, MI), you’ll join us in person one day every other month (6 times per year) for a fun, purpose-driven Collaboration Day. These days are filled with creative energy, meaningful connection, and team brainstorming!

Must be able to work during Eastern Time unless approved by your manager.

Employees working remotely must have a dedicated, ergonomically appropriate workspace free from distractions with a mobile device that allows for productive and efficient conduct of business.