Shipping open data to standard: CTDL credentials and FHIR health data

Open standards only mean something if someone does the unglamorous work of getting data into them and keeping it accurate. This case study is about that work: the data operation, the credential publishing, and the engineering direction, across two domains. Workforce credentials in New Jersey, and health data for CMS Medicaid. Why open, portable public data matters is argued elsewhere. Here the question is how it got built and kept running.

Publishing New Jersey's credentials to CTDL

On the workforce side, Chelsea led the hands-on work of publishing New Jersey's training and credential data to CTDL, the Credential Transparency Description Language maintained by Credential Engine. CTDL is a national, linked-open-data vocabulary. Describing the state's programs in it means they read the same way as credentials anywhere else in the country, rather than sitting in a New Jersey-only format. Credential Engine now lists MyCareer.NJ among the data published to that standard.

Hitting a national standard is not a one-time export. It meant mapping the state's real program records to CTDL's vocabulary, reconciling them against wage-record outcomes, and folding the work into the same pipelines that keep the public Training Explorer current. The catalog residents see is the catalog that feeds the CTDL pipeline. So the data published in an open, machine-readable form, the kind other agencies and systems can actually consume, stays in step with the public site.

The data operation behind it

Standards-based publishing is only as good as the data feeding it. So the underlying data operation is the real engineering story. The catalog of 3,800+ approved training programs is kept accurate every week and matched to real wage-record outcomes. It has run for more than three years without a lapse.

• 3,800+ approved programs reconciled and kept current, not a static snapshot
• Listings kept current, so they match the real catalog
• Weekly accuracy held for more than three years, with no gap in coverage
• The catalog that powers the public site also feeds the CTDL pipeline

Leading the FHIR specialization for CMS Medicaid

On the health side, Chelsea leads Coforma's specialization in FHIR and health-data interoperability, setting the technical direction so federal Medicaid systems can exchange data in standard, modern ways. It is the same open, linked-data move as CTDL, applied to clinical and program data rather than credentials. Her teams own engineering across the CMS Medicaid data-collection suite: seven open-source repositories in a public Enterprise CMCS organization, built on AWS CDK micro-services with Lambda, DynamoDB, API Gateway, Kafka change-data-capture, and Cognito for identity, on a React and TypeScript front end using the CMS Design System.

This is leadership, not personal hands-on coding. The direction is hers; the delivery is carried by named feature leads and the teams under her. In one stretch a brand-new application went from zero to production in about fourteen weeks, and the portfolio's 2026 delivery moved through hundreds of merged pull requests, the output of a team she leads, building interoperable health data in the open.

Why deliver it this way

Publishing to CTDL and FHIR is more work than shipping a private format, and the payoff lands downstream. Agencies can coordinate, data can be reused, and a resident is not asked to re-prove something they have already earned. What runs through both portfolios is delivery discipline: open standards adopted in practice, on a data operation reliable enough that the standard actually means something.