NEW EPISODES. REAL STORIES. DELIVERED TO YOUR INBOX.

Join listeners navigating the space between one chapter and the next.

Katie Savage
The most important thing any technology organization can do is build in-house expertise rather than outsource its thinking to vendors.
Katie Savage
← PrevWork Unscripted · S02 E14Next →
Coming Soon

This episode hasn't dropped yet. Subscribe to get notified when it's out.

Subscribe for Updates →
Work Unscripted

She Started as CIO of Maryland 8 Months Pregnant. Here's What She Built.

with Katie Savage

🎧SpotifyYouTube

Katie Savage took the CIO job for the state of Maryland eight months pregnant. She came from Defense Digital Service, where she learned that the most important thing any technology organization can do is build in-house expertise rather than outsource its thinking to vendors. In Maryland, she found a team that managed contracts. She built a different kind of team.

Key Takeaways

  • A Leadership Coach Told Her to Stop Leading Like a Man: When Katie was stepping into the DDS director role, her leadership coach told her that if she tried to lead like her male predecessors it would not feel natural and she would fail. That permission to lead through her own strengths, specifically listening, collaborating, playing politics, and building teams, was the moment something clicked and she found her groove.
  • Maryland's CIO Office Owns Its Own Fiber Network and Radio Grid: Unlike most state CIO offices that focus primarily on contract management, Katie's remit includes running Maryland's own fiber internet network, one of the only state-owned networks in the country, and the radio network that all first responders in Maryland use. When the January 2025 DC plane crash occurred, many of those calls were made over her team's radio infrastructure.
  • Estonia Showed Her That Applications May Become Obsolete: During a trip to Estonia, Katie asked her hosts how citizens apply for benefits and was met with confusion because Estonia simply looks at the data and delivers the benefit automatically without requiring an application. She sees AI enabling the same shift: instead of building application interfaces, teams will focus on the underlying data sources and models that trigger automatic eligibility.
  • She Recruited Former USDS and DOGE Staff to Maryland: Because so many digital service veterans in the DC metro area were looking for work after federal cuts, Katie was able to bring a wave of former USDSers and DOGE employees into the Maryland state government to staff her newly created digital service team. She recruits primarily on mission rather than salary, offering opportunities like protecting critical infrastructure and responding to live ransomware attacks.
  • Peace Corps Taught Human-Centered Design Before It Had a Name: Katie spent two years in Malawi with Peace Corps and credits a Brown spiral-bound handbook called the Participatory and Capacity Assessment handbook with teaching her the skills of following people in their day-to-day lives, presenting information back to communities, and understanding problems before jumping to solutions, which she later recognized as the foundation of human-centered design.

In This Episode

  • What it was like to start a major CIO role eight months pregnant
  • How she shifted Maryland's tech organization from vendor management to in-house expertise
  • What Maryland's infrastructure actually includes: their own fiber network, first responder radio, and the public benefits platform
  • What female leadership in tech looks like when you stop trying to lead like your male predecessors

What We Discuss

How she got to DDS and why the CIO track was always the plan
Starting as Maryland CIO eight months pregnant and what that was like
What it means to lead as a woman and stop trying to lead like your male predecessors
Building in-house talent and recruiting former USDS employees
The public benefits platform: SNAP, WIC, and what it means to serve six million people

Q&A

Questions answered in this episode

What skills do you need to become a state CIO?

Katie says you need a solid mix of IT ops, which covers networking, cybersecurity, enterprise platforms, and identity and access management, combined with product skills like design, engineering, and understanding user needs. She also increasingly emphasizes data proficiency, specifically the ability to collect, manage, analyze, and display data, as a foundation that cuts across all of it.

How does a state government compete with private industry for tech talent?

Katie says Maryland cannot compete on salary and she is candid about that, but she recruits on the specificity of the mission: offering cyber candidates work on live ransomware attacks, 24/7 security operations center experience, and critical infrastructure protection. She also credits Governor Moore's reputation as a forward-looking leader as doing a significant portion of the recruiting work for her.

What is the biggest challenge of consolidating state IT under one CIO?

Katie says the core problem is that government agencies think about services by agency rather than by the citizen's outcome, so a person trying to start a business gets linked out to seven different departments instead of seeing a single list of required permits. She is also working to reverse decades of fragmentation where agencies bought their own tools independently, including 13 separate Microsoft contracts she found when she arrived.

How do you balance innovation with military or government culture?

Katie describes it as diametrically opposed pressures: you are there to disrupt, but you are operating inside a culture intentionally built around discipline, chain of command, and tradition. Her approach was to spend her early time at DDS in deep listening mode while simultaneously introducing operational changes she knew were needed, rather than coming in with solutions before she understood the environment.

What is the future of product managers and UX designers in an AI world?

Katie believes the roles will not disappear but the tasks will look completely different, shifting away from coding dropdown menus toward understanding what information people need from an AI-powered interface and how to translate that into data queries. She compares it to the Estonia model where the goal is not to build a better application but to eliminate the need for one by surfacing the right data automatically.

More Conversations

Keep Listening

Every episode is a different story about the space between one chapter and the next.

Browse All EpisodesSubmit a Guest