In late 2015 I joined Studio 0815 — a design studio working on flight and mission systems for defense — as its first employee. Nine years later I was directing a department of 3 teams, and I never stopped shipping. Below: the roles that arc covers, and three of the systems my teams and I designed for the Israeli Air Force.
The Israeli Air Force's mission plan and control application, intended to serve as its main tactical tool: a wide overview of the whole military picture — air, ground, ocean, underground — from the soldier in the field to the whole campaign, plus command and control of aerial forces.
The client had also asked for a separate system to plan aerial missions before they're sent to the control app. I folded the two into one platform, so planning and controlling are performed by the same system and each user can focus on their specialty within it.
Each one pulled in a different direction, and the system had to answer all three at once — before a single screen could be considered settled.
A wide range of user types and digital-literacy levels, from the Air Force commander to the intelligence clerk in the squadron.
The system had to cover a wide range of actions and present the relevant data in each view.
It had to work with the IAF's other systems and terminology.
All fluent Hebrew speakers, and very unlikely to be colorblind. Beyond that they diverge completely — and the difference that mattered most was leverage: some were a captive audience for the system, others could easily replace it with something else.
It was clear early that the system had to be map-based. The harder problem was defining the core entity — the mission — since every user saw it, and defined it, from a different perspective.
So it got taken apart. Everything anyone called part of a mission was collected, then sorted until it stopped resisting: two kinds of thing, each with its own small set of verbs. That split became the layers menu, the tools rail and the contacts menu — the analysis is the navigation.
The full set walks the system from its core surface through the menus and panels that support it, and ends where the process actually ended — on a sketch that was expected to change.
Coloring and styling this system was a challenge on its own — not only because of the amount of detail, but because of the customization required for its many different users. One example: users have five categories for classifying an aircraft, multiplied across every type of force a user might need to view.
The result is a full design system — icon library, components, dark and light variants — built to hold that complexity.
Buttercup plans, commissions and manages reconnaissance missions — photography and recording missions performed by drones — for the Israeli Air Force, the Intelligence Corps and the General Staff.
It replaced a workflow run on Excel, PowerPoint and email.
Two problems they called main, two they called secondary. The design answered the main pair directly — and, as it turned out, got the secondary pair as a side effect of the same structure.
Commissioning missions and avoiding conflicts was a tedious job: there were too many variables to consider. Main.
It was almost impossible to change plans or reassign missions in real time. Main.
Every action needed approval by multiple people, which was hard to keep up with. Secondary.
The organization needed to document the trail of changes and the discussion history. Secondary.
Intelligence asks, the Air Force plans, the General Staff reviews. Daily, daily, weekly — and the weekly reviewer is the one with the least patience for the interface.
Research combined user interviews, on-site observations, and analysis of the artifacts officers were already using — their Excel sheets, emails and PowerPoint decks. That surfaced a taxonomy of the existing process, built around Platform and Mission, which I reframed as four questions: What, Who, When, and How-and-where.
The questions resolved into a new taxonomy — Request, Platform, Time, Mission — bound together by a new Reconnaissance event, which also carried the approval chain and captured the officers' dialogue, solving the secondary problems along the way.
Selected from the full flow: the missions bank and the timeline that bind to it, the map the officer plans on, and — in the design's own annotations — what each of Buttercup's two officer roles actually needed from it.
My initial approach used a light look-and-feel, unlike the traditional dark, high-contrast army style — I believed dark, high-contrast designs put users under stress and led to hasty decisions.
The beta group disagreed. I moved.
A light interface, against the operational convention, on the argument that high contrast under pressure produces hasty decisions.
They found the light design uncomfortable, and felt it didn't read as an operational system or convey the consequences of their actions.
A dark interface, with a light timeline acting as a focus box over the main work area.
The client opened this project with a list. Closing it means answering that list, item by item, rather than declaring the project a success.
Commissioning and real-time reassignment both resolve into the Reconnaissance event: one entity to create, move and reassign instead of a set of coordinated artifacts.
The approval chain rides on the event, which is what made batch approval possible on the timeline.
Solved by improving the dialogue between the system's users: a logging tool designed as a chat, so every event and action could be commented on and stay part of the record.
The Israeli Air Force's squadron management system, built to serve every squadron type — combat, transport, helicopter and drone.
Many officials plan the squadron's long- and short-term agenda together, across operational, recreational, private and collective matters.
Squadron management means addressing operational and personal matters together, for every squadron type at once — and the bar the new system had to clear was set by what the squadrons were actually using instead of the old one.
Combat, transport, helicopter and drone squadrons, on one system.
Many officials plan the agenda side by side, across operational, recreational, private and collective matters.
The system Bubbles replaced was army-made and cumbersome; in practice, the work happened in Excel spreadsheets, Outlook email and calendar, and erasable whiteboards.
Bubbles's design problem is the range of its users: a deputy commander planning six months out, a reservist checking one week bi-weekly, an ops clerk living in it all day. The same screen has to serve someone with zero time to learn it and someone who resorts to Excel at any inconvenience.
Research meant visiting the squadrons, interviewing many users, and learning their Excel sheets and documents directly. It became clear the system to design was a calendar with multiple perspectives: four types of calendar event represented on it, with each screen serving every user but acting as the main workspace for some.
Eight pages from the v1 flow: the two grids every view resolves into — month and week — the two ways to read a week (staffing, or a Gantt bar), and the panels a day's work actually runs through — a new event, the inbox, the day itself.
The first stage set the working surface — the Gantt, switchable between day, week and month. The redesign that followed is where the visual system got built: dozens of icons drawn, fonts and colors selected, on a light theme that kept the color code users were already used to — the one piece of the old workflow worth carrying forward.
The shipped (v2) generation, walked scale by scale — month, week, staffing, Gantt, day — with the design's own callouts naming which of the six personas each screen is a primary workspace for, and which a secondary one.
Three case studies show the work. The rest of the job was building the organization that could keep producing it.