Fusion Twin: Designing a Scientific Simulation Platform

Overview

Key Achievements

• Gold Winner — Vega Digital Awards 2024 (Science & Technology).
• Bronze Winner — Summit International Awards 2025.
• The interface demonstrated scalability in a $10B+ market, helping secure key R&D funding.

Understanding the Scientific Workflow

Preparing data, configuring simulations, and exploring results required switching tools. The fragmented workflow raised a question:

How can we bring these steps into one clear experience?

First, I needed to understand the users

Experienced researchers needed detailed controls, reliable data, and flexibility. Students needed guidance and a safer way to experiment. The challenge was supporting expert work without overwhelming beginners or limiting access to scientific detail.

Research showed where the biggest friction was

Interviews and workflow analysis revealed:

The same problems appeared again and again:

The opportunity was not to simplify the science.

Organizing the information architecture

I grouped features around researcher tasks, connecting data, setup, and analysis in a clearer information architecture.

Choosing what to build first

I focused on four core tasks:

The resulting priorities:

• file management
• simulation setup
• status tracking
• results exploration.

Exploring the workflows before the final UI

Early sketches explored layout, hierarchy, and actions before visual detail.

Wireframes tested complex workflows through:

This resolved structural issues before visual design.

Wireframes validated parameter grouping and system feedback.

Core Solutions

Giving researchers one place to start

The dashboard brings files, simulations, saved outputs, and visualizations together, making current work visible at a glance.

Data became the foundation

Researchers work with complex datasets, including HDF5 files. I designed one workspace to upload, organize, inspect, and reuse data across simulations, bringing file management into the product instead of leaving it outside the workflow.

Simulations needed more guidance

Simulation setup involves many parameters and technical decisions. I designed a guided sequence that makes required steps clear to beginners while keeping advanced controls available to experts. This balanced guidance with the flexibility scientific work demands.

Managing many simulations

Researchers needed to track simultaneous simulations by state:

• draft
• queued
• running
• completed
• failed.

Clear states let users track progress without opening each simulation.

Understanding relationships between parameters

Parameter mapping makes relationships between files, parameters, and outputs visible and configurable before analysis.

The result was only useful if users could understand it

Researchers needed to compare experimental data, simulation outputs, and multiple runs. Configurable chart layouts bring these datasets into one workspace, making differences easier to explore without switching between separate tools.

Designing scientific data changed how I thought about simplicity

Hiding scientific information would undermine the work. I organized complexity instead: essential information appears first, while detailed data and deeper controls remain available when needed. This approach supports both quick orientation and thorough analysis.

Building a foundation for future features

To support new scientific workflows, I created reusable patterns for:

• tables
• forms
• file states
• simulation statuses
• graphs
• spacing
• responsive behavior.

Shared patterns gave design and development a consistent foundation.

The result

What I learned

Fusion Twin taught me that a complex product does not always need to become simple.

The more useful question is: Which complexity belongs to the science — and which complexity is created by the tool?