Engineering leader for public systems. I build teams, portfolios, and the technical foundations that let them deliver.

Most recently, I was a Senior Director of Engineering at Coforma, one of three Senior Directors in a 50-person engineering organization. I led a 22-person reporting structure with five direct reports, including three engineering Directors, owned the company-wide healthcare engineering portfolio, and served as engineering business-development lead.

That work stayed close to outcomes:

• Nine engineers moved into senior or leadership roles, including four promotions to Director or Principal. My teams had three voluntary departures across three years.
• I supported $150M+ in 2024 federal contract wins at a 70%+ win rate and held delivery responsibility behind $100M+ in awarded prime contracts.
• I owned engineering for the seven-application CMS Medicaid & CHIP Data Collection Tools suite, oversaw Coforma's work on the federal Medicaid Drug Programs system, and led the firm's FHIR and health-data-interoperability specialization.
• I was engineering lead and principal engineer for MyCareer.NJ.gov, New Jersey's statewide workforce platform. It has served 1.8 million users since December 2023 and provides bilingual career, training, labor-market, tuition-assistance, and support-service tools.

On MyCareer.NJ.gov, I led the platform while continuing to set architecture across its three production codebases. The work included a zero-downtime GCP-to-AWS migration, test coverage raised from zero to 93–96%, a 94% reduction in known vulnerabilities, complete English and Spanish access, and CTDL publishing that made New Jersey training credentials interoperable with the national Credential Engine ecosystem.

I also designed its shared applied-AI foundation, now live in production, and six primary proof-of-concept features that remained gated from end users pending pilot approval. The most mature was a bilingual, grounded career coach with cost-aware model routing, a merge-blocking adversarial test gate, and evaluation designed to expose weak categories instead of hiding them. The production platform served 1.8 million users; the AI features were POCs, not deployed services.

Before Coforma, I built public systems from inside California government:

• At the California Energy Commission, I led development of regulatory energy-data systems, helped move the commission's data infrastructure to AWS, and operated the Title 20 appliance certification system.
• At the California Public Utilities Commission, I directed the Transportation Carrier Portal launched in 2021 and architected a claims portal for telecommunications-access and broadband programs.
• At the California Department of Social Services, I built and operated caregiver-safety, licensing-payment, legislative-analysis, and benefits-program systems.
• At UC Berkeley's Graduate School of Education, I led application development for the equity-focused 21st Century California School Leadership Academy.

I hold an M.S. in Software Engineering from CSU Fullerton, completed while working full-time in state government, a B.S. in Computer Science from the University of Oregon, and an ITIL Foundation certification.

Based in Davis, California · Portfolio · LinkedIn

These are independent projects that are public, inspectable, and ready to name. Most are pre-1.0 tools or reference implementations, not evidence of organizational adoption. Each repository states what is shipped, what remains experimental, and what still needs human review.

• tods-validate is a deterministic validator for the Transit Operational Data Standard. Version 0.8.0 is published on PyPI and ships a GitHub Action, container image, pre-commit hook, and browser playground.
• GTFS Scorecard is a live transit-data-quality service covering a curated registry of 1,700+ feed records and 1,100+ published scorecards. It also ships a GitHub Marketplace Action and a read-only MCP server.
• NearMiss is a beta road-safety analysis toolkit that uses exposure-normalized rates, confidence intervals, and statistically controlled hotspot detection rather than presenting raw report density as risk.
• Swelter is an alpha community heat and air-quality reference system with correction provenance, a bilingual dashboard, and OGC SensorThings exports.
• ledger is a beta, privacy-first digital-preservation tool using BagIt, PREMIS, Dublin Core, encrypted contributor identities, and consent-based disclosure. It contains synthetic and consented fixtures only.

I lead at portfolio altitude and stay technically deep enough to set architecture, challenge assumptions, and understand delivery risk. Grit is not a staffing model: I build clear ownership, strong managers, useful standards, and systems that do not depend on one heroic person.

My current technical center of gravity is TypeScript, Python, React and Next.js, AWS, PostgreSQL, data interoperability, applied AI evaluation, and delivery systems. Earlier work also spans C# and .NET, Salesforce, Snowflake, Azure, GCP, and Natural/ADABAS modernization.

I treat accessibility, privacy, security, operability, and multilingual delivery as engineering requirements. For AI systems, that means evidence-constrained generation, explicit refusal, adversarial testing, disaggregated evaluation, cost controls, feature flags, and human review for legal, policy, subject-matter, and accessibility judgments.

AI agents are part of my development workflow. I choose the architecture, write the acceptance criteria, review the output, and decide whether it is ready to release.

• I won't work on weapons or warfare, policing or mass surveillance, or technology that profits from incarceration. I also won't work with organizations based in, or strongly tied to, Israel.
• I look for organizations whose work helps people routinely failed by public systems, and whose leadership reflects the communities they serve.
• I won't use AI to decide whether someone gets a job, benefit, service, or opportunity. It can support a person's decision, but that person should be able to inspect the evidence, correct bad information, and make the final call.
• I collect as little personal data as practical and prefer local or offline tools when they fit.
• I separate sources, calculations, and interpretation. When a system does not know, it should say so.

MyCareer.NJ.gov has been featured by Fast Company and the Royal Statistical Society's Real World Data Science. The Harvard Kennedy School's Project on Workforce described its Training Explorer as a "promising potential model," and Credential Engine has featured its CTDL publishing work. The platform was also nominated for a NASWA Workforce Innovation Award. In 2023, I joined a Digital Services Coalition round table on AI and machine learning in federal and civic technology.

I'm interested in W-2 engineering leadership roles such as VP of Engineering, Head of Engineering, Senior Director, Director, or Principal Engineering Manager. I want to lead people and managers, shape architecture and delivery systems, and build reliable, accessible technology in public health, workforce and social services, energy and utilities, government digital services, or responsible AI.

Reach me through chelseakr.com or LinkedIn.