Independent product design & engineering studio
Make the angle where insight meets object.
Tangent Works is a fictional studio for hardware-adjacent software, interfaces, devices, and systems designed for people whose work is messy, constrained, and real.
01
Not a template. A test rig.
A studio page usually promises polish before showing process. Tangent Works starts closer to the bench: constraints, prototypes, failure modes, and the line from idea to object.
We design like engineers and build like designers: drawings that can fail, code that can be felt, and materials that answer back.
-
Insight before interface
We map constraints, workarounds, physical context, and real use before choosing a screen, component, or material.
-
Prototype as argument
The fastest way to settle a debate is an object, flow, fixture, or interaction that people can touch and test.
-
Production in the sketch
Specs, interface logic, mechanical assumptions, and handoff notes are created together, so delivery is not a cliff.
02
Specimen sheet
Capabilities arranged like material samples: each tab shows a different part of the studio line from research to release.
Map the problem before naming the solution
Research becomes a shared coordinate system: who is using the product, where they are, what breaks, and what cannot be changed.
- Constraint mapping and workflow archaeology
- User interviews with technical, operational, and regulatory partners
- Material, platform, and dependency scans
- Visible risk registers and open questions
Design as a set of decisions
Screens, states, physical forms, and interaction rules are treated as proposals, not decoration. Each one should survive being clicked, carried, interrupted, and argued with.
- Interaction flows and low-fidelity prototypes
- Component libraries for complex interfaces
- Accessibility review across contrast, focus, motion, and input methods
- Hardware-aware service design and wayfinding
Engineering that keeps the idea testable
We build the working line: interface logic, device integration, data paths, firmware assumptions, and the instrumentation needed to learn while things are changing.
- Product interface architecture and state management
- Firmware, device APIs, sensor plumbing, and service integration
- Test harnesses, fixtures, and simulated conditions
- Prototypes built for evaluation, not applause
Production readiness as an experience
The goal is not a final Figma file. It is a maintainable system: documented, componentized, reviewed, and handed over to people who will keep it alive.
- Pilot support and observation of real use
- Release notes, decision logs, and technical handoff
- Design systems with ownership paths
- Component cleanup before production pressure arrives
03
Selected explorations
These are fictional, illustrative studio studies. They demonstrate design thinking and system shape, not client names, launches, or measurable results.
-
Field maintenance companion
A ruggedized mobile checklist study for service technicians working with noisy environments, gloves, interrupted attention, and shared devices.
-
Museum wayfinding label
A signage and audio-label system designed for calm navigation: adaptive content, clear focus, and paths that respect accessibility constraints.
-
Clinical intake tablet
An intake workflow study that assumes interruption, shared devices, privacy needs, and staff who cannot afford confusing states.
04
Engagement
A fictional studio works best in short arcs. The goal is not infinite discovery; it is producing the next useful object and the next honest question.
-
Frame
Start with constraints, failure modes, users, existing tools, and the smallest useful promise. No broad transformation narrative, no decorative research theater.
-
Prototype
Build a working line: interface, mechanism, data path, fixture, or interaction. The prototype should create better arguments than slide decks alone.
-
Harden
Prepare documentation, test notes, component rules, and handover artifacts for teams who will maintain the product after the studio leaves.
Next
Send the mess.
The most useful first message is not polished. Share the constraints, users, interruptions, failure points, and what you already believe. Then we can find the angle.