Skip to main content

Data & Analytics

Find out how we turn government data into decisions people can trust, and the skills you need at each level.

Introduction

Data & Analytics practitioners help teams and clients make decisions grounded in evidence. They build the pipelines, platforms, and models that turn raw government data into something usable, and they translate analytical findings into insights that move products, services, and policy forward. They work across the full data lifecycle — collection, cleaning, storage, analysis, transformation, visualization, and communication — while protecting privacy and security.

Core craft

Data pipelines and platforms

Building, operating, or working with the systems that ingest, transform, store, and serve data reliably

Modeling and analysis

Selecting and applying the right methods — from simple aggregations to advanced models — to solve the problem

Code quality and engineering practices

Writing readable, tested, reusable code that other practitioners can understand and extend (for Data Engineers, production pipeline and platform code; for Data Scientists, reproducible analysis code and notebooks others can rerun and build on)

Security, privacy, and compliance

Protecting sensitive data — including personally identifiable information (PII) and protected health information (PHI) — across the full lifecycle, from ingestion through analysis and presentation

Communication and storytelling

Translating data work into insights that stakeholders can understand and act on (for Data Engineers, making the system legible to technical partners; for Data Scientists, turning insights and tradeoffs into decisions for non-technical stakeholders)

Stakeholder and mission grounding

Working with stakeholders to identify the right questions, and connecting data work to the client's mission outcomes

Associate

Develop foundational data engineering or data science skills while supporting a delivery team.

Scope: what you decide and own
Executes defined data engineering or analysis tasks, with guidance
Relationships: working with stakeholders
Responds to stakeholder requests with guidance; contributes to shared artifacts like pipeline code, notebooks, queries, and visualizations
Reach: what changes because of you
Their assigned data work is done reliably, with little rework — pipelines run, queries return correct results, and analyses and visualizations are clear and ready to use
Calibration question
Can they deliver defined data tasks reliably, with guidance?
What success looks like
  • Supports pipeline, analysis, or visualization work with reliable follow-through
  • Writes clear, working code or queries that match the team's conventions
  • Handles sensitive data carefully and follows privacy and security guidance
  • Communicates findings and technical decisions clearly to teammates
  • Asks questions to understand the data, the stakeholder, and the underlying program
Example outcomes
  • A pipeline component or query that runs correctly and matches the team's conventions
  • Clean, well-documented code that another practitioner can pick up
  • A chart or dashboard panel that answers a specific stakeholder question
  • A handoff of de-identified or synthetic data set up correctly for development and testing

Level I

Apply data engineering or data science methods independently to a defined piece of the team's data work.

Scope: what you decide and own
Decides the approach for well-scoped data work; owns a defined piece of a larger data effort — a pipeline component, a specific analysis, a dashboard, or a transformation
Relationships: working with stakeholders
Builds solid working relationships with their direct counterparts on the team; communicates findings and technical decisions clearly
Reach: what changes because of you
The data work they own is done well and on pace
Calibration question
Can they choose the right approach for well-scoped data engineering or analysis work without close oversight?
What success looks like
  • Owns a pipeline component, a specific analysis, a dashboard, or a transformation within a larger data effort
  • Writes working, tested code with minimal guidance
  • Picks appropriate standard methods for the problem — aggregations, joins, regressions, clustering, basic visualizations
  • Identifies data quality or pipeline issues early and resolves or escalates them
  • Builds direct relationships with the team members and stakeholders who depend on their work
  • Articulates the value of their work to teammates and stakeholders
Example outcomes
  • A pipeline transformation that handles a tricky source-system quirk without breaking downstream
  • An analysis that gave the team a clear, defensible answer to a stakeholder question
  • A dashboard that the team and client actually use to track program outcomes
  • A migration of a small dataset off a legacy system with no loss of fidelity (Legacy Modernization variant)
  • A metric definition or data collection approach that gave the team a defensible way to measure a program outcome
  • A code review on a teammate's pipeline that caught a real issue

Level II

Lead a pipeline, analysis, or model end-to-end through ambiguity.

Scope: what you decide and own
Decides the approach for ambiguous data problems; owns a pipeline, an analysis, or a model end to end (e.g., a multi-source ingestion pipeline for a Data Engineer, an end-to-end analysis or model for a Data Scientist)
Relationships: working with stakeholders
Manages stakeholder expectations proactively on data and analysis questions; facilitates working sessions and reviews that end in decisions; produces client-ready data products, analyses, and visualizations that land without heavy revision
Reach: what changes because of you
The pipeline, analysis, or model they own delivers its intended outcome — it runs reliably, the analysis answers the question, the model performs in production
Calibration question
Can they own an ambiguous pipeline, analysis, or model end to end and make the calls without escalating?
What success looks like
  • Owns a complex pipeline, an end-to-end analysis, a model, or a measurement problem end-to-end — including framing the question and defining the metrics, not just building
  • Defines and tracks the outcome the work is meant to produce — not just whether it ran
  • Makes methodology and architecture tradeoffs and explains the reasoning clearly
  • Runs the stakeholder conversation for their work without senior backup, including pushing back on questions the data can't actually answer
  • Produces client-ready artifacts — analyses, models, visualizations, pipeline documentation — that land without heavy revision
  • Handles privacy and security tradeoffs appropriately, including de-identification choices and re-identification risk
  • For a Data Engineer, may own a substantial pipeline or platform component end-to-end; for a Data Scientist, may own an end-to-end analysis or production model
Example outcomes
  • A multi-source ingestion pipeline that consolidated data from three legacy systems into a single, query-ready store
  • An analysis that surfaced a previously unseen pattern in eligibility, enrollment, or program participation and shifted how the team prioritized work
  • A production model — predictive, classification, or otherwise — that performs reliably and is interpretable enough for stakeholders to trust
  • A dashboard or visualization that became the canonical view the client uses for a program metric
  • A problem-framing and metric-definition effort that changed what a program was measuring, not just how
  • A privacy review and de-identification approach for a sensitive dataset that let the team move forward without compromising protection

Senior

Own the data platform or analytical capability for a product or program and guide its direction.

Scope: what you decide and own
Makes the consequential tradeoff calls on architecture, methodology, and how data is used; owns a data platform or analytical capability their operating context — their engagement or program — depends on
Relationships: working with stakeholders
Navigates competing priorities and turns messy data situations in their area into clear choices for decision-makers; runs the data conversation with the client for their product or program without senior backup
Reach: what changes because of you
Their area's data outcomes — what gets built, how it performs, what decisions it enables — improve because of decisions they made, not just work they delivered
Calibration question
If they stepped away, would the team lose data judgment, not just capacity?
What success looks like
  • Owns the data platform, pipeline architecture, or analytical capability for a product or program of meaningful scope
  • Makes consequential tradeoff calls on architecture, tooling, methodology, and/or how data is used to inform decisions
  • Acts as a trusted advisor to client stakeholders, including senior ones, on data questions and reframes problems for them
  • Runs the full data conversation for their product or program without senior backup
  • Anticipates data, privacy, or compliance risks before they hit the team
  • Builds alignment across product, design, engineering, and the client around how data flows through the system and what it's used for
  • Mentors junior and mid-level practitioners on the team through code review, working sessions, and direct collaboration
  • Strengthens the team's data and analytical judgment alongside their own work
  • Contributes to business development on the engagement — recompetes, scope expansions, or proposal input on data work
Example outcomes
  • A data platform redesign that cut pipeline runtime substantially while reducing cost, and that other teams started borrowing from
  • An analytical capability built into a product that gave program staff a real-time view of an outcome they couldn't previously see
  • A modeling approach adopted by the client across multiple programs after evidence of impact on a first program
  • A privacy and compliance approach for sensitive data — including de-identification, access controls, and audit — that let the engagement scale up data sharing safely
  • A first-time technical lead on the team who's now running their own data workstream
  • A renewed or expanded contract grounded in data work the client trusts

Staff

Improve data outcomes across the whole operating context.

Scope: what you decide and own
Sets data direction for their whole operating context and owns the systems and standards it runs on — shared data platforms, modeling standards, privacy and compliance approaches, cross-team data architecture
Relationships: working with stakeholders
Gets teams across their operating context pulling in the same direction on data priorities and influences program-level data strategy with client leadership; clears data blockers no single team can clear
Reach: what changes because of you
Their whole operating context produces better data work because of the systems, standards, and direction they put in place, and it travels to other teams and engagements
Calibration question
If they stepped away, would data work across their operating context lose direction, not just a strong contributor?
What success looks like
  • Sets data direction across multiple workstreams or teams
  • Identifies systemic data issues — quality, architecture, methodology, privacy — across engagements that no single team would see
  • Establishes shared practices — data platforms, modeling standards, privacy and compliance approaches, code review norms — that other engagements adopt
  • Resolves tradeoffs around data architecture, tooling, or analytical approach across the operating context
  • Mentors practitioners across teams or engagements
  • Influences program-level data strategy with senior client stakeholders across more than one program area
  • Strengthens cross-team decision-making on data and analytical questions
  • Contributes to Skylight's growth through proposal contributions, scope expansions, or new opportunity development
Example outcomes
  • A shared data platform or analytical framework adopted across multiple engagements that reduced ramp-up time for new teams
  • A modeling or analytical standard — interpretability requirements, evaluation criteria, documentation expectations — that improved consistency across the practice area
  • A cross-engagement privacy and de-identification approach that gave Skylight a defensible position when negotiating data access with new clients
  • A synthesis across two or three engagements that revealed a shared client data challenge and shifted how Skylight approaches that domain
  • A new contract or scope expansion that came directly from data work and relationships they built
  • A practitioner who's now making complex tradeoff decisions because of their coaching

Principal

Shape data strategy across portfolios or mission areas, and help open new opportunities for Skylight.

Scope: what you decide and own
Decides, with client leadership, what data should accomplish in part of the client's mission, cutting through the noise to diagnose the root cause; shares ownership of Skylight's own health and growth
Relationships: working with stakeholders
A trusted advisor up to client leadership on data strategy, analytics, and how data informs the mission — aligns stakeholders on the solution and earns their buy-in; a recognized voice in the field who represents Skylight externally in data engineering, data science, or the client domain
Reach: what changes because of you
The client approaches part of its mission differently because of their data work, and Skylight is stronger for it — they help open new opportunities through proposals, expanded scope, or new clients
Calibration question
If they'd never been here, would the client's approach to its mission — or Skylight's growth trajectory in data — be on a different path?
What success looks like
  • Reframes problems and influences how the client thinks about data's role in part of its mission
  • Aligns client leadership around data direction across multiple programs — what to collect, what to model, what to make actionable
  • Guides major tradeoffs on data architecture, analytical investment, or how data informs policy and program design
  • Acts as an external authority on data engineering or data science in the client's domain — through speaking, writing, or recognized expertise
  • Contributes to Skylight's Data & Analytics practice area through hiring, leveling, craft standards, and supporting practitioners across the practice area
  • Contributes to Skylight's growth through proposal contributions, expanded scope with existing clients, or new client opportunities
  • Improves organizational learning from how data work runs across engagements
Example outcomes
  • A program or agency function reorganized around a data approach the Principal brought — changing how the client measures outcomes, allocates resources, or designs services
  • A new contract won at an existing client because of data credibility built across prior engagements
  • A proposal effort they led or contributed to that brought in new mission work for Skylight, particularly work that required deep data expertise
  • A data architecture or analytical approach adopted across a client portfolio that materially changed how the agency uses data
  • A change across the whole practice area at Skylight — a leveling rubric, a privacy standard, a modeling approach, a hiring rubric — that raised the bar across all Data & Analytics engagements
  • An external presence (talks, publications, recognized expertise) that brought new opportunities to Skylight

We’re hiring!

Let’s work together to change the way government
serves millions of people.