Jackson DeVico

they/them

Human-Computer Interaction Master's student - University of Maryland
B.S. Games and Simulation Arts and Sciences - Rensselaer Polytechnic Institute

Demonstrating waste-water treatment to environmental engineering students through 360 simulation

Experimental Media and Performing Arts Center (EMPAC)

January 2024 - April 2025

Overview

This project involved creating a 3D simulation of a waste-water treatment plant for use as a supplement to a field trip of an actual waste-water treatment plant.For the first year of development, I was the sole developer of the project under the advisement of a professor and an EMPAC research engineer. After this year, I recruited a group of students, and my role shifted towards management.

The first demo of the waste-water treatment plant

Problem

The target audience was undergraduate students taking an introductory environmental engineering course. The professor's proposal sought a virtual model of a waste-water treatment plant that would achieve the following:

The initial time between starting development and the first in-class demonstration was four months, after which the future direction of the project would be evaluated. Later on, a (potable) water treatment simulation was also developed, after which a larger team of students was formed to begin a migration to Unreal Engine.

Process

After meeting with the professor to discuss project needs, I met with the EMPAC research engineer to discuss what technology we would move forward with. We decided that Unity Engine and Blender were ideal for modeling and rendering the simulation because work for the 360 screen using Unity Engine already existed, and Blender is a free established software that would allow for easy transitions between personnel in the future.The project went through four development segments:

For the first three phases of development, the process involved creating 3D models, texturing the models, integrating the models into Unity Engine, and adding details such as animation, particle effects, physics, and camera setups. Every two weeks I would meet with the professor to receive feedback on how to improve the realism of the models and adapt them to the specific lecture the professor intended to give.

One issue I encountered was how to properly display tall objects on the 360 screen. While the screen is tall, the much larger circumference results in a very wide aspect ratio, cutting off taller objects (as seen above). To account for this, I created models with walls removed, and added additional camera angles that bring the entire object into view.

For the last few months I was involved in the project, I recruited students with more specific specialties to handle the most significant portions of development as we began to migrate the simulation to Unreal Engine. We used Notion to track team progress, and my responsibilities shifted towards managing team efforts, running meetings, providing progress reports, and creating documentation.

Results

By the end of my time on the project, I had developed a functional simulation of both a waste-water treatment plant and a water treatment plant. Features include multiple camera angles (including cross-sections), pipe flow visualization, and physics-based particle movement simulation. Alongside a team, work was also started on an Unreal Engine version of the simulation, which would have more realistic lighting and textures, more detailed models, and audio.After each demonstration, the professor collected feedback from students through direct questions and take-home surveys. The results consistently showed that students felt they had a better understanding of water treatment processes after experiencing an immersive lecture with the simulation.The success of the project also helped jumpstart an initiative from other professors to collaborate with EMPAC to develop similar teaching tools. I worked on two other similar projects for shorter lengths of time, one which combined Google Maps data with drone footage to show rock layers for a geology class, and another which used 3D scans to teach people about how cleanrooms are used.

Updating the Visual Design of "Testudo"

University of Maryland - INST 711: Visual Design Studio

September 2025 - December 2025

Overview

Testudo is the Registrar’s website for the University of Maryland. Students use the website to register for classes, view their schedules, access other sections of UMD’s website, and more. However, the visual design of the site is outdated, broken, or does not support users.In this project we update the visual design of Testudo to improve the usability and visual appeal of the site. We limit our focus to four pages: the home page, schedule of classes, student schedule, and final exams.While involved in most aspects of this project, the aspects in which I was solely responsible for were the heuristic and accessibility evaluation, design library, and Schedule of Classes pages.

Problem

Testudo’s visual design of the site is outdated, broken, or does not support users. Our goal is to update Testudo’s visual design. The timeline for the project is 6 weeks. Our process is as follows:

Process

Current Design Analysis

Testudo's existing home page

The existing design of Testudo has many issues, including:

Heuristic Evaluation

A heuristic evaluation of Testudo's pages of interest revealed significant issues. Below is a brief list of the Nielsen Norman heuristics that are relevant to each screen. A detailed written audit can be found here.Home page:

Class registration page:

Schedule of classes/final exams pages

Contrast Evaluation

Design Inspiration

Our visual design direction is based on two sources. The first is UMD’s existing visual branding. We chose to utilize Testudo’s shell logo because it is simple, modern, and can fit in many locations. The second is the website for Marathon, a game developed by Bungie.

On top: UMD branding, on bottom: Marathon website

Low Fidelity

Our low-fidelity stage focused on exploring our visual direction and blending visual styles into a cohesive design vision. We explored typography and layout decisions for multiple screens. This screen uses a card layout that we expand upon in mid and high-fidelity.

Mid Fidelity

The mid-fidelity schedule of classes screen expands upon the low-fidelity version by adding relevant information to the cards, and moving more towards the Marathon visual design language. The search options are also condensed to be simpler to scan and interact with. We would continue to use progressive disclosure as a design principle moving forward.

High Fidelity

Moving into high-fidelity, we began bringing our designs together to promote cohesion. We unified our color choices (UMD’s base colors + accent colors for links and schedules) We also altered our grid systems to not use gutters, and instead use thin border strokes to separate sections.

We also introduced a “light mode” concept, and created some alternative screens for light and dark mode in our finishes product.

Design Library

Our design library evolved throughout the design process, but the following is a representation of our core design elements, colors, and typography.

Results

Our final designs blend a modern, minimalist, technological aesthetic with the University of Maryland’s iconic Testudo branding and colors. The new pages emphasize simple information architecture, progressive disclosure, visual elements with meaning, and a unified, consistent design language across all the pages. Moving forward we would expand upon our screens to implement light mode and dark mode for all of them, conduct user research and collect feedback, and iterate based on said feedback.

Schedule of classes, desktop, dark mode

Schedule of classes, desktop, light mode

A Contextual inquiry of student
engagement and utilization of GenAI summarization tools in an academic context

University of Maryland - INST 710: User Experience Research Methods

September 2025 - December 2025

Overview

This summary is non-exhaustive. A detailed report of the entire project can be found hereThis project involved examining student use of a generative AI (genAI) summarization tool as a lens to learn about their specific wants, needs, and expectations of genAI in an academic context.Project responsibilities were divided fairly evenly, with all members involved in processes such as affinity diagramming and interviewing. In addition, I was allocated the role of project manager, and I managed scheduling, organized meeting, and maintained project software (Notion).

Problem

Our research questions were the following:

The timeline for the project was about three months, and consisted of identifying the problem statement and research methods, data collection, analyzing and creating models, and proposing design interventions. The research culminated in an in-class wall walk at the end of the semester.

Process

Contextual Interviews

We utilized a mixed-methods approach to collect specific insights while still reaching a large population. Our primary research method was contextual interviews, and our secondary research method was digital surveys.We conducted five in-person contextual interviews with students from University of Maryland's Human-Computer Interaction Master's program. These interviews were all approximately one hour long, and had three phases:

Surveys

Surveys were open to the entire University of Maryland student population. Students were recruited mostly through announcements in digital spaces and in-class announcements from professors. The survey had three sections:

Affinity Diagram

After each data collection, we held interpretation sessions to align our team understanding of the data. Using our findings from these sessions, we created an affinity diagram.

Sensation Board

As a team, we divided to create additional models to supplement the affinity diagram. I worked on the sensation board, which is a model designed to capture the experiences and sensations that an ideal product should manifest. Our sensation board captures four core ideas: optimized flow, in control, confident clarity, and conscientiously engaged.

Design Visioning

Based on our observations and models, we creating design visions that implement our findings. We started with three visions, evaluated those visions, and ultimately conceptualized one cohesive design intervention.Through our final vision “WorkSpace”, we delivered a central, customizable AI learning environment that supports optimal flow, confident clarity, and a strong sense of control through adjustable settings, guided refinement tools, and transparent citation modes.

Results

Our key findings were that:

During a wall walk, we presented our models, design concepts, and findings, and received feedback from observers. Many people left notes on our affinity diagram expressing that they agreed with certain sentiments from our affinity diagram, especially regarding ethical concerns surrounding AI usage. Many visitors also left notes asking potential follow-up questions to our research.