Work / SaaS · CAD / One-week challenge

3D Modeling UI

One week to design a browser-based 3D modeling interface for mechanical and industrial engineers, a professional tool with an expansive viewport, built for users who need advanced functionality, not casual hobbyists. The constraint made the method matter more.

Role
Solo designer, research, concept, interaction, UI in one week
Domain
SaaS · CAD / 3D parametric design
Methods
Kanban, user interviews, desk research, persona, empathy map, A/B testing, heuristic evaluation
Audience
Mechanical & industrial engineers
Context
One-week design challenge
Outcome
A research-backed CAD interface, built in one week
Browser-based 3D modeling interface: an expansive viewport with a model on the grid, a tool rail at the left, and a parameter panel at the right
The interface: viewport first, tools at the rail, parameters where the eyes already are.
Challenge

Pro-grade CAD, in a browser, in a week

The brief: craft an interface for a 3D modeling software delivered as Software-as-a-Service via web browsers, empowering users to create intricate 3D elements. I focused not just on meeting the provided specifications but on conceptualizing the design’s structure, detailing the interactions, and making the thought process behind each decision explicit.

Two commitments framed the work: design for users who require advanced functionality rather than casual hobbyists, and guard the software’s capacity for future enhancement: in features and platform compatibility, so the design accommodates evolving user needs and technological advancement.

Process

A week run on Kanban

To monitor progression I ran the project on the Kanban agile methodology, a fluid, adaptable workflow for a compressed timeline. Within the week, the design thinking process still encompassed a full series of interviews, research, user studies, evaluations, and A/B testing: user-centric and data-informed, with iterative improvements and strategic decision-making to the last day.

Kanban User interviews Desk research Persona Empathy map A/B testing
Kanban board with To Do, In Progress, and Completed columns tracking research and design tasks
The week’s Kanban: research, persona, empathy map, and evaluation moving across the board.
Research

Five interviews, one literature

Constrained by time, I conducted five user interviews, and engineered around the limitation: a question guide of open-ended questions enabled semi-structured interviews, conducted via video call under pandemic conditions, each recorded with explicit participant consent strictly for research purposes. Ethics aren’t a luxury feature of fast research.

Desk research anchored the interviews in the CAD usability literature: working on three-dimensional layouts through a two-dimensional screen is one of the classic interface problems in engineering tools. Following Lee et al., the top CAD UI issues range across dialog boxes, drawings and reports, modeling, domain terminology, view and navigation, menus and toolbars, and their three families of principles (general system design, 3D-parametric-specific, and user support) later structured my evaluation. Notably, Nielsen Norman Group’s ten heuristics serve simple desktop and mobile software well, but do not stretch to the interface of a sophisticated 3D system.

Pie chart of reported CAD interface issues: Modeling 26%, Dialog box 25%, Terminology 17%, Help 8%, Drawing and report 8%, Menu items 5%, Toolbar 5%, Others 4%, View and navigation 2%
Where CAD interfaces fail users, from the literature: modeling, dialog boxes, and terminology lead the complaints.
Persona

Meet Fabian

The research converged on one person: Fabian, 35, a mechatronics engineer in Köln who makes at a local hackerspace one day a week. He moves between Inventor, SolidWorks, SketchUp, and Fusion, picks up new skills from YouTube, and loses time hunting through crowded menus for the one tool he needs. His goals, behaviours, pain points, and motivations kept the design honest: built for a capable professional with very little time to waste.

Persona sheet for Fabian, a 35-year-old mechatronics engineer in Köln: bio, behaviour, goals, pain points, personality, and motivation
The persona: Fabian, the capable, time-poor maker the interface is built for.
Four-quadrant empathy map: says, thinks, feels, does, built around the engineer persona
The empathy map: says, thinks, feels, does, summarizing the persona and qualitative data.
Design

From persona to viewport

The interviews produced an engineer persona and an empathy map split into its four quadrants, a glimpse into who the user is as a whole, keeping the design user-centered rather than feature-centered. The interface that followed is guided by fundamental design principles: an expansive modeling viewport at the center, a consistent tool rail, parameter controls beside the model tree, and terminology that matches the engineer’s domain rather than the developer’s.

The design was then evaluated against the same principles the literature provided, closing the loop between desk research and the finished screens.

The default 3D modeling workspace: viewport, tool rail, and parameter panel
The default workspace, model centered in the viewport.
The same workspace with the projects rail expanded on the left, showing project thumbnails and import or export controls
The projects rail expanded: switch files, import, and export without leaving the canvas.
One week is not enough time to skip the research. It’s exactly why the research had to be sharper.
process reflection