Defense politics
How ideology and domestic politics drive military action, from budgets and procurement in Washington to the nationalist movements shaping Russia's war in Ukraine.
- What ideas justify a war at home?
- Who shapes what comes after?
Open to product design roles in defense tech
Designing for
Designing for
decisions.
I'm Julian Zeppa, a product designer, co-founder of Armada Strategic, and political science student at the University of Miami. I design and ship software, from government intelligence tools to consumer apps, and I'm happiest when the problem involves real users and messy data.
ProfileJZ-01
JZ
Julian ZeppaProduct designer · Co-founder
- Company
- Armada Strategic
- School
- University of Miami, 2027
- Focus
- Defense & modern warfare
- Shipped
- Civic.OS · Guppo
Design2 products shippedA government platform and a consumer iOS app
CompanyCo-founder, Armada StrategicDefense technology & government contracting
EducationB.A. Political ScienceUniversity of Miami · Class of 2027
How I work
-
01
Understand
Go where the work happens
Before I designed anything for a district office, I spent a summer working in one, answering constituent calls and tracking cases by hand. The pain points were obvious from the inside.
In practice SF District 2 → Civic.OS -
02
Structure
Make the important thing obvious
Civic.OS turns five data sources into three questions staff ask every morning. Guppo turns 144 species into one collection you can read at a glance.
In practice Civic.OS · Guppo -
03
Ship
Build it, then keep going
I take ideas all the way to working software, building with AI-assisted development. Civic.OS is in development for a real office, and Guppo is live on the App Store.
In practice 2 products shipped in 2026
Featured work
All work →
Armada Strategic · Government
Civic.OS
Designing an intelligence platform that fuses five data sources into one evidence-backed daily briefing.
Read the case study →
Independent · iOS
Guppo
A consumer iOS app: photo-based species ID, a collectible 144-species Dex, and privacy-first maps.
See the project →Designing software for defense and national security.
Get in touchAbout
I'm a product designer and co-founder of Armada Strategic, and I'm finishing a B.A. in political science at the University of Miami. Most of my classes and research focus on defense: how nations decide to build and use military power, and how technology is changing what war looks like.
I didn't take a straight line here. I've worked in a supervisor's office in San Francisco, in law firms in Miami and Los Angeles, at the polls on election day, and behind the counter of a dive shop. Each one taught me how people actually get work done under pressure. I want to bring that, along with what I've learned about defense, to the software analysts and operators rely on.
- Studying
- B.A. Political Science, University of Miami, 2027
- Focus
- Defense politics, Russia & Ukraine, maritime security
- Building
- Armada Strategic · Civic.OS · Guppo
- Based
- San Francisco · Miami
Research focus
University of Miami · Political ScienceWarfare & security
How violence starts and spreads: insurgency and guerrilla tactics, state collapse, non-state actors and maritime security, and what modern conflicts teach about drones, autonomy and how fast militaries adapt.
- Why does violence emerge where it does?
- Can it be predicted before it happens?
Defense technology
The shift toward software-defined defense: new entrants in the industrial base, data and decision-support tools, and how the people using them stay in control.
- How do operators trust what software tells them?
- Where does the human stay in the loop?
Selected research
Research assistant
I work as a research assistant to Professor Costantino Pischedda at the University of Miami, studying how guerrilla insurgencies use conventional and unconventional tactics.
The 2009 Hijacking of the Maersk Alabama
When Desperation Meets Opportunity
A case study of the hijacking of the Maersk Alabama and the capture of Captain Richard Phillips, set against Somalia's collapse after 1991. I argue that piracy isn't random. It's predictable when two conditions meet: opportunity, where weak central governance leaves the sea without rule of law or deterrence, and motivation, where people lose the livelihoods that kept them out of it.
To test that, the paper compares Somalia with piracy in the 17th- and 18th-century Caribbean. There, distant empires left whole territories ungoverned, and the end of privateering in 1713 left skilled seamen without lawful work. Two theatres, centuries apart, show the same pattern.
- Question
- What conditions make maritime piracy develop?
- Method
- Case study and historical comparison
- Frameworks
- Weber's state monopoly on force; economic violence
- Finding
- Piracy follows predictable conditions: permissive environment plus loss of livelihood
Full paper available on requestRequest a copy →
It's All About the Blood
A Deep Dive Into Russian Ultranationalism
An analysis of Russian ethnonationalism and its role in the war in Ukraine. I trace how the Kremlin's Russkiy Mir ideology, reinforced by Alexander Dugin and the Russian Orthodox Church, turned the idea of a "diasporic empire" into a justification for annexing Crimea and invading Ukraine.
The paper then compares two armed nationalist groups fighting on opposite sides: the Russian Volunteer Corps, which fights for Ukraine and wants a smaller, ethnically homogeneous Russia, and the Russian Imperial Movement, a U.S.-designated terrorist group that wants an even more expansionist one. I close by weighing which scenario is more likely in a post-Putin Russia, and why the outcome of the war decides it.
- Question
- What role could ethnonationalists play in a post-Putin Russia?
- Method
- Ideological analysis and comparison of two armed groups
- Topics
- Russkiy Mir, Dugin, the Orthodox Church, RVC, RIM
- Finding
- Both groups can sway Russia's future, and the war's outcome decides which
Full paper available on requestRequest a copy →
Selected coursework
- POL 571Prof. McGuinn · 2026 · Maritime piracy case study
- POL 384Prof. Beck · Spring 2025 · Russian nationalism and the war in Ukraine
Why design for defense
I've spent years studying how militaries and governments make decisions. Good tools help people make those decisions faster and more safely, with the evidence in front of them. That's the software I want to design.
Work
02 projects
01 / 02 · Armada Strategic · Built for SF District 2
Civic.OS
An intelligence platform for a city council district office. It fuses 311 tickets, the office phone log, local news and Reddit into one operating picture, and each morning answers what is happening, what changed and what needs attention. Every finding links back to the records behind it, and every figure comes from a stated rule.
- Role
- Product design lead; built with AI-assisted development
- Stack
- React, TypeScript, Express, Supabase, Leaflet
- Year
- 2026
- Status
- In development for SF District 2 · live demo
02 / 02 · Independent
Guppo
An iPhone app that turns fishing into a collection. Snap a photo of a catch, Guppo identifies the species and unlocks its card in a 144-species Guppodex, with personal bests, weekly challenges and a catch map that keeps exact spots private by default.
- Role
- Product designer and creator; built with AI-assisted development
- Stack
- React Native, Expo, TypeScript, Supabase, SQLite
- Year
- 2026
- Status
- Live on the App Store
Civic.OS
Giving a city council office one clear picture of its district every morning. After I pitched it, the supervisor asked me to keep building it for his office.
I found the problem from the inside
In summer 2026 I worked as a legislative intern in San Francisco's District 2 supervisor's office. A district office is the first place residents call when the city lets them down, but staff see the district through four disconnected windows: a phone log in a spreadsheet, 311 tickets in the city's system, local news, and what neighbors post online.
Nobody had time to connect them. So the questions that mattered went unanswered: which problems are growing, which ones the city marked "fixed" without fixing, which departments are falling behind, and where the office can actually make a difference today.
Phone logSpreadsheet
311 ticketsCity system
Local newsOutlets & RSS
NeighborsReddit, Nextdoor
The office already had a mental model. I built around it.
When I asked how staff decide what to act on, a supervisor described two distinctions. Is a problem acute, a one-time occurrence, or general, continuous and ongoing? And for acute problems, is it actionable, something the office can do now, or does it call for outreach, telling residents with no immediate action?
That became the Triage view. Rather than invent a new scoring system, I turned the office's own decision tree into explicit rules, so the software sorts issues the way staff already think about them.
I worked alongside the whole office, four staff and the supervisor, for the full summer. The biggest finding was simple: staff spent hours entering data by hand, and almost none of it was ever analyzed. The work was tedious, and the records it produced weren't answering anyone's questions.
"Is it a one-time thing or does it keep happening? And can we actually do something about it today?"
The two questions behind every triage decision, paraphrased from the office
How might we tell a district office what's happening, what changed, and what needs attention every morning, without giving staff more work?
What I changed, and why
Rank by open work, not by novelty
Before
The first briefing led with an 11× spike: 11 reports on one street. Dramatic, but small.
After
The real story was a 3× rise with 52 reports still open. I reweighted priority by open workload, with the multiplier capped.
Readable rules over a black box
Risk
Staff won't act on a flag they can't explain to their boss or a resident.
Decision
Everything that ranks or flags is a stated rule. Each issue shows why it surfaced and links to its records.
Measure from the data's edge, not the clock
Before
City data publishes a day late, and police reports trickle in for days, so every neighborhood looked quiet.
After
Trends count back from the newest published record and leave out days still being filed.
Be strict about "our district"
Before
Long streets cross district lines, and a supervisor can share a surname with a better-known politician.
After
A story only counts if it names a district neighborhood, landmark or corridor. A street name alone isn't enough.
Humans stay in charge
Risk
Software speaking for an elected official.
Decision
Civic.OS drafts escalation letters but never sends them. Staff review and send from their own email.
Only public information goes to AI
Risk
Constituent calls and case notes leaking to an outside model.
Decision
AI only summarizes published news and public posts. Private data is blocked in two separate places in the code.
One operating picture, built from evidence
Trust was a design requirement, not a feature
| Risk | How the design handles it |
|---|---|
| Internal data reaching an outside AI | Public sources only, enforced twice. Every AI overview is labeled and linked to its source. |
| Residents' contact details | Masked in lists, excluded from reports, and every view is written to an append-only audit log. |
| Who can do what | Individual accounts with viewer, operator and admin roles. |
| Speaking for an official | Drafts only. Staff always send. |
What happened
I pitched Civic.OS to the supervisor at the end of my internship. He asked me to keep developing it for the office, and I still am.
What I learned
Data products earn trust by explaining themselves. Showing the rule behind each flag and the records behind each sentence mattered more than any single feature.
In progress
Adding tools to track local politics and to monitor continuing issues in the district over time, so problems that keep coming back stay in view.
Julian Zeppa
Product designer · Co-founder, Armada Strategic · B.A. Political Science, University of Miami (2027)
Experience
-
Mar 2026 – PresentSF Bay Area
Co-Founder · Strategy, Operations & Product Design
Armada Strategic LLC- Co-founded a defense and security company pursuing government contracting. Led SAM.gov registration, federal contractor onboarding, procurement and NAICS/PSC research, capability statements and market-entry plans.
- Led product design for Civic.OS, from user research inside a district office to shipped software: an operations dashboard, an evidence-linked daily briefing, pattern detection, maps, escalation tracking and an audit log.
- Designed the information architecture around three questions staff ask every morning: what's happening, what changed, and what needs attention.
- Pitched Civic.OS to the District 2 supervisor, who asked me to keep developing it for his office.
- Produced Armada's brand, video ads and the Civic.OS case study.
-
Aug 2026 – PresentIndependent
Product Designer & Creator
Guppo (iOS)- Designed and launched a consumer iOS app, now on the App Store, with photo-based species ID, a collectible 144-species Dex, weekly challenges and a social feed.
- Designed privacy by default: friends only ever see an approximate catch area, never the exact spot.
-
May 2026 – PresentSan Francisco
Legislative Intern
City and County of San Francisco · Office of Supervisor Stephen Sherrill, District 2- Handled constituent services and outreach for the district office.
- Mapped how staff tracked constituent issues across calls, 311 and email. That research became the basis for Civic.OS.
- Helped create and run District 2 entertainment zones, including the Union Street Entertainment Zone and World Cup events.
-
University of MiamiResearch
Research Assistant
Prof. Costantino Pischedda, Department of Political Science- Research on the use of conventional and unconventional tactics by guerrilla insurgencies.
-
Nov 2025 – May 2026Miami
Legal Intern
Altonaga Law PLLC- Researched and drafted appellate briefs and memoranda for federal criminal cases.
-
May 2025 – May 2026Miami
Strategic Partnership Chair
Zeta Beta Tau Fraternity- Ran sponsorship outreach by email and phone, and managed the lead pipeline.
- Built repeat partnerships with brands including Peet's Coffee and Red Bull.
- Managed the budget and logistics for social, philanthropy and brotherhood events.
-
Jun – Aug 2025Los Angeles
Legal Assistant
Wisner Baum- Managed records for more than 5,000 mass-tort settlement claims in FileVine and ProLaw.
Earlier
- Summer Sales Associate, Bamboo Reef Scuba Diving Centers 2024 – 2025
- Poll Worker, City and County of San Francisco, multiple elections 2021 – 2023
- Summer Legal Intern, Moore Law Group, mass-tort data management 2022
Education
2023 – 2027
University of Miami
B.A. Political Science, Minor in History · GPA 3.52. Provost Honor Roll & Dean's List (Fall 2024, Fall 2025, Spring 2026). Focus on defense politics, international relations and modern warfare.
2019 – 2023
Stuart Hall High School
High school diploma and International Baccalaureate Diploma.
Skills
Product design
User researchInformation architectureInteraction designData visualizationPrototyping
Building
AI-assisted developmentReactReact NativeTypeScriptSupabaseNetlify
Domain
Government operationsDefense & gov contractingInternational relationsElections
Research & business
Legal researchEvidence reviewPartnershipsBrand & video
Contact
Let's build something that matters.
jvz9@miami.edu