WAT Researcher
Desktop-First Text Analytics
Role
Sole UX Designer & Engineer
Duration
6 months
Team
Me (UX designer and Engineer ), Project Manager, Principal Investigator
A case study in strategic product pivoting, describing the design & redesign and engineering of a desktop-first academic NLP tool to replace an aging desktop tool.
TL;DR
The initial plan was to build a web app for a critical research tool, but it became clear this approach would be too costly and complex. I proposed and led the shift to a desktop application instead.
I handled the entire process—from UX strategy and design to full-stack engineering—building the platform with Angular, Django, and Electron.
This decision fixed major scaling problems, provided a much richer feature set for thousands of researchers, and gave the organization a new capability in desktop development.
NDA limits what I can show, but I've included a few design shots, commit logs, and IA—happy to share more over a call.

Challenge
-
Legacy System Replacement: The Learning Engineering Institute (LEI) required a modern, web-based alternative to Coh-Metrix, an outdated yet widely-used research tool serving over 5,000 researchers worldwide.
-
Front-end Completion and Backend Roadblock: Although the front-end was successfully built using Angular, the backend text-analysis component proved extremely resource-intensive. Hosting this analysis as a web-based service would lead to significant operational expenses and scaling difficulties.
-
Scalability and Cost Risks: Due to the high computational demands of backend processing, the project's feasibility and sustainability faced critical risks, jeopardizing the overall success and continuation of the initiative.

Role & Contributions
As the sole UX Designer and Engineer, my role expanded to cover key strategic and technical responsibilities.
-
Proposed the Pivot to Desktop: When it became clear the web-first approach wasn't viable, I researched and presented the case for switching to a desktop application, solving a core business and technical problem.
-
Designed the New Architecture: I designed the complete user experience and technical architecture for the desktop app. This included a new project-based workflow that was a major improvement over the original web concept.
-
Built the Full-Stack Application: I developed the entire application, handling the Angular frontend, the Django/Python backend, and packaging it for both macOS and Windows with Electron and PyInstaller.
-
Delivered the Original Web App: I also built the initial 3-page web interface. After we shifted focus, we repurposed it as a complementary tool for quick, simple analyses.
-
Mentored the Team: I am currently leading junior team members, guiding them in creating user documentation and running our QA testing plan.

Process
-
Initial Web App Delivery: I initially designed and built the first version of the web interface, which met our initial milestone at LEI and served as a proof of my front-end skills.
-
Identifying the Core Problem: During backend integration, I flagged the high server costs, queuing complications, and technical hurdles we would face, which started the discussion about finding a better way forward.
-
Proposing the Desktop Solution: I developed a proof-of-concept for a desktop app to show it could handle more complex features (like project management and local files) and solve the scalability issue.
-
Designing a Better Experience: Once the new direction was approved, I designed a more complete user experience based on a researcher's actual workflow, adding features for project creation, managing large sets of documents, and a local task queue.
-
Building and Delivering: I handled the entire desktop development process, including the challenge of cross-platform packaging with Electron and PyInstaller to create installers for macOS and Windows.

Solution
The Solution: A Desktop and Web App
The final result is a solution with two parts, each for a different type of user:
-
WAT Researcher (Desktop): The main application for in-depth research. It's a powerful tool that lets users manage projects, handle large amounts of data, and run complex analyses on their own computer.
-
WAT Researcher (Web): A simple, lightweight version for quick analysis of a single text or small document set. It offers easy access to the core features with no installation needed.

Outcomes and Impact
This project is scheduled to launch and replace an essential tool for a large academic community.
-
Strategic Impact: Changing from a costly web architecture solved a major business challenge and will save the institute significantly on server and maintenance costs.
-
Organizational Impact: I built the institute's first-ever pipeline for developing and distributing a desktop application, a completely new capability for the organization.
-
Product Impact: We delivered a product that is more powerful and versatile than originally planned, better meeting the needs of both casual and advanced users.
-
Career Impact: This project was a key step for me into a role with more strategic leadership and mentorship, proving I can guide a technical product from concept to launch.
Tech Stack
Tech & Methods
- Frontend: Angular, HTML5, CSS3
- Backend: Django, Python, SQLite
- Desktop: Electron, PyInstaller, Shell Scripting
- Design: Figma, Wireframing, Prototyping, User Flow Mapping
- Process: Agile Development, Stakeholder Management, Technical Strategy
Growth
This project shows my ability to lead a complex, multi-platform product from concept to launch, combining the strategic thinking and technical skill required for senior UX Designer or UX Engineer roles.