About Skylight

Skylight is a digital consultancy using design and technology to help government agencies deliver better public services.

We’re at the forefront of a civic movement to reinvent how all levels of government serve families, patients, and many others in today’s digital world.

If you want to play a part in driving this critical movement forward, we’d love for you to join our growing team of public interest technologists.

The work we do matters.

About the job

At Skylight, data scientists organize, synthesize, visualize, and analyze data to help government stakeholders and product teams make informed decisions. They apply a variety of analytical models to find meaningful insights in both structured and unstructured data, and they communicate these insights in an easily digestible format.

The Administration for Children and Families (ACF) is expanding surge support and capacity for its new Data Capacity and Digital Services organization within the Office of Administration, through the option year of our Task Order 1 contract. The team works with enterprise-serving organizations like ACF Tech and the Office of Grants Management to help programs deliver on ACF’s mission, including priorities focused on reducing grantee burden; reducing waste, fraud, and abuse; and modernizing legacy systems.

You’ll help answer hard questions with data. This role also flexes into lightweight engineering: when a problem calls for a small tool or automation built with AI, DataBricks, or Power Apps rather than a model, you build it.

What to know going in

This work calls for a consultant’s skill set: you’ll interface directly with senior leaders, and much of the work isn’t scoped in advance. This isn’t a team where you take tickets off the top of a backlog. Alongside delivery, we do a lot of technical consulting and capacity building, so people who take initiative and are comfortable with uncertainty and ambiguity tend to do best here.

While the scope of this role is likely to evolve over time, the majority of this work may not focus directly on predictive analytics or statistical analysis. Rather, the majority of your work would be focused on data ingestion, data storage and processing, and data visualization.

What you’ll do

Engage executive and program stakeholders to frame the right questions, then take the initiative to go find the answers, even when the ask is ambiguous or the audience is tough
Organize data and build data pipelines, using tools like Palantir Foundry and PowerBI
Build AI agents, integrating them into workflows, validating outputs
Visualize grant and program data using Tableau and PowerBI
Analyze large amounts of structured and unstructured information, including grants and program data, to discover trends and patterns
Build lightweight tools, prototypes, and automations with platforms like Credal and Power Apps when a quick solution beats a long one
Propose solutions and strategies for business challenges, and select the right tools, frameworks, and languages for the job, with a preference for open-source, and low/no-code solutions
Communicate findings through data visualization that leadership can act on
Build predictive models and machine-learning algorithms, at times, and
Collaborate with others as part of a cross-functional team that includes user experience researchers and designers, product managers, engineers, and other functional specialists

What we’re looking for

Basic qualifications

Can think about and use data strategically to identify the right questions to ask, communicate ideas, surface opportunities to create business value and drive decisions
Familiar with modern data engineering, including techniques for cleaning and preprocessing data from a variety of sources
Ability to analyze a data set using the appropriate analytical method and can design, prototype, and build statistical models that are credible, effective, and accessible
Ability to preserve privacy and security when performing data analysis
Can articulately communicate your data, models, and findings to technical and non-technical audiences