Remote, On Campus

Undergraduate Students, Masters Students

WeAudit - Supporting the Crowd in Auditing AI and Algorithmic Systems

Software Developers and UX designers

Undergraduate Students, Masters Students

People are already organically coming together to audit algorithmic systems for bias and harm. The goal of WeAudit is to build tools that can help support everyday people in this task, with a focus on generative AI. Our tools include image comparisons (e.g. one could see "professors" is mostly male and "teachers" is mostly female), and basic application of computer vision to make it easy for people to see how things like age, gender, and skin color vary based on terms fed into generative AI.

We're looking for software developers and UX designers to help us in building new tools and supporting existing ones. See our website at https://weaudit.org/ for examples.

Yiya AirScience: Delivering STEM Education through Radio-based and Phone-based Technologies for Learners in Uganda

Remote, On Campus

aeo@andrew.cmu.edu (Amy Ogan), jstamper@cmu.edu (John Stamper), ckwon2@andrew.cmu.edu

Yiya AirScience is a remote learning model that provides radio-based and phone-based virtual classrooms to deliver innovative engineering education for learners in rural Uganda. Many learners reside in low-infrastructure contexts, in which they face numerous obstacles, including lack of educational resources, expensive schooling fees, lack of internet access, etc that prevent them from receiving basic educational experiences. As radios and keypad phones are the most commonly possessed technologies in Uganda, the remote learning model aims to eliminate these barriers by providing unique STEM learning experiences through a radio-based and phone-based learning system. As multiple courses from Yiya AirScience have been deployed, our research aims to investigate student learning through data to better understand and predict course performance and persistence. We are looking for student researchers interested in investigating diverse learning analytics for student learning in these remote courses.

Undergraduate Students, Masters Students

We are developing Negotium, an interactive mobile application for learning about and practicing everyday negotiation skills, using a human-centered design approach. Research shows women negotiate less often than men across a variety of situations and this tendency perpetuates inequality such as the leadership and wage gap between men and women. One way to work towards equity is by teaching and providing scalable access to training tools that will empower women to advocate on behalf of themselves.

We are looking for a qualified UX designer to join the Negotium team. We've implemented a first version of the application, and this spring with your help we will user test the application and iterate on and extend the design. In this project, we will work with PROGRESS (Program for Research and Outreach on Gender Equity in Society) a CMU founded organization which addresses gender inequity by providing negotiation tools to empower women and girls.

Students should be able to spend 10-20 hours per week on the project and attend weekly project meetings. This position is available for pay or independent study credit.

Interested students should send their resume, major, GPA, and links any relevant projects to dabbish@cmu.edu"

Remote, On Campus

Undergraduate Students, Masters Students

There are standard tests for how fast a person can use a pointing device like a touchpad or mouse, and how fast a person can type. However, there has never been a standard way to measure how fast people can scroll. We developed a prototype test, and used it in a class, but a number of problems were identified with it. The project for Spring'2024 is to improve the test, and if there is time, run more user tests with the new version. The goal is to write a new paper to submit to a conference based on the new test.

Undergraduate Students, Masters Students, PhD Students

People with limited functional motion may rely on orthotics to perform daily activities or aid in rehabilitation. This project aims to understand how users interact with and perceive assistive orthotics. By evaluating customized orthotics through both short-term and longitudinal studies, we will provide guidance the future design and development of personalized orthotics.

Task description:
Short-Term Study: Measure and evaluate the effort savings and enhancements provided by orthotics.
Long-Term Study: Assess comfort, satisfaction, and user usage of orthotics over an extended period.

Remote, On Campus

Undergraduate Students, Masters Students

Mosaic is an Extensible Framework for Linking Databases and Interactive Views. Ad part of this project, you will develop new algorithms, interfaces, and evaluations for this framework used by hundreds of developers worldwide to create lightning fast machine learning and data science applications.

https://github.com/uwdata/mosaic/issues?q=sort%3Aupdated-desc+is%3Aissue+is%3Aopen+label%3Aprojects

Undergraduate Students, Masters Students, PhD Students

Over the past several years, there has been increased investment in automated vehicle (AV) technology for full-size public transit buses, with deployments planned across the country. Yet, operating transit is more complex than the light-duty passenger vehicles for which AV technologies are commonly designed. Buses are significantly larger and operate in highly variable environments with many stops and starts near vulnerable road users (VRUs) such as pedestrians and cyclists. Furthermore, transit operations require supporting passengers and maintaining safety inside the vehicle. Due to both technical and operational challenges, transit vehicles including buses and vans will continue to require skilled human drivers, even as automated vehicle capabilities are incorporated.

To help maintain transit’s already high level of safety for passengers, it is essential to understand how automation stands to affect the roles and day-to-day tasks of trained drivers. For example, driver assistance automation such as pedestrian warnings and lane-centering has the potential to improve the safety and workload of drivers. At the same time, automation can create new kinds of safety issues caused by the interactions in human-autonomy teams and can intensify work as people primarily take over from automation in the most challenging situations. Therefore, it is crucial to carefully consider how to go about incorporating automation technologies into fleets and to develop new interfaces for transit drivers to work effectively in maintaining safety on the road, and alongside vulnerable road users.

In this research, we draw on our ongoing partnerships with two major transit unions to develop technologies and interfaces that would help bus drivers in their work, and improve operations. This work will also include passenger-facing design engagement through partnerships with local rider advocacy organizations focused on understanding how these technologies could enhance overall rider experience, with the aim of supporting safety and accessibility. Over the course of the semester, we will conduct workshops and other participatory design activities to imagine and develop co-designed prototypes of new automation technologies and human-machine interfaces to support a more just and accessible future of transit for drivers and passengers.

Remote, On Campus

Undergraduate Students, Masters Students, PhD Students

We are looking for a student to contribute to developing and testing web applications that leverage fine-grain learner performance data from tutoring systems. The application is designed to help students learn to set goals and stay motivated to achieve them.

This position is ideal for students looking to augment their portfolio or tap into research. Involvement in research activities, such as through UX tests and academic publications, is subject to the applicant's initiative and interest.

https://ies.ed.gov/funding/grantsearch/details.asp?ID=5810

Undergraduate Students, Masters Students, PhD Students

Teaching volunteers and professionals the therapeutic skills needed to help clients deal with mental health and behavioral problems is expensive and slow. In addition to receiving didactic materials about the nature of the skills they need, trainees need opportunities to practice these therapeutic skills and receive feedback. This project explores the use of large-language models like ChatGPT to engage trainees in synthetic conversations in realistic conversations with a virtual patient (i.e., a client chatbot) without putting real patients at risk. Our goal is to refine and evaluate a working prototype of our current client chatbot. We will work closely with volunteer support-providers from the online peer-to-peer therapeutic service 7Cups.com or students in counseling programs.

Students who are interested in the project will have the opportunity to participate in the following research activities: (1) conducting formative interviews and observational studies to identify the skills trainees need most; (2) building and testing a virtual patient generator to specify the type of patient with whom trainees will converse, in terms psychological problem (e.g., job stress, depression, loneliness), personal style (e.g., open or resistant to revealing personal information), or background (e.g., teenage girl, middle-aged male employee); (3) building and testing a module for providing feedback to the trainee; and (4) conducting qualitative or formal user studies to understand how the conversations with the chatbot can be best used to improve training.

Undergraduate Students, Masters Students

PolicyCraft is a CSCW system designed to facilitate collaborative and participatory policy development. We've developed the first fully functional version and conducted an initial study. Before proceeding with further studies in partnership with additional online communities and local governments, we plan to refine the current version of PolicyCraft. We are seeking students with a strong background in web development to assist with this iteration. Students can also run subsequent studies together, co-author research papers, and work with Prof. Ken Holstein, Prof. Haiyi Zhu, and our collaborators in other institutes.

HCI Research Intern - Physical Prototyping

Undergraduate Students, Masters Students

This independent study explores the use of mobile robots (Sony Toio) with custom-designed attachments to address smart device privacy. For example, what if a robot could carry a small mechanical hand to block the view of a smart camera. The goal is to design and build five functional simple prototypes inspired by My Little Piece of Privacy (https://www.niklasroy.com/project/88/my-little-piece-of-privacy) and Project Alias (https://github.com/bjoernkarmann/project_alias)

Remote, On Campus

Undergraduate Students, Masters Students, PhD Students

Welcome for Web3 degens, users and analyst! We aims to explore the social and psychological dynamics behind Meme Coins and NFTs within the Web3 community, especially focus on Sui, but Solana and other blockchain are welcome as well. Through a combination of qualitative interviews and quantitative data analysis, the project seeks to understand the motivations, values, and strategies of community members, including traders, holders, developers, and promoters. The findings will provide valuable insights for the Web3 Blockchain Foundation and the broader blockchain community, with the goal of fostering a more resilient ecosystem. Please send me your CV and your experiences in Web3 by Email.

https://x.com/Eason_C13/status/1854254898524692782

Yiya AirScience: Delivering STEM Education through Radio-based and Phone-based Technologies for Learners in Uganda

Remote, On Campus

Undergraduate Students, Masters Students

Yiya AirScience is a remote learning model that provides radio-based and phone-based virtual classrooms to deliver innovative engineering education for learners in rural Uganda. Many learners reside in low-infrastructure contexts, in which they face numerous obstacles, including lack of educational resources, expensive schooling fees, lack of internet access, etc that prevent them from receiving basic educational experiences. Many African countries, including Uganda, are also heavily influenced by their colonial history, which is reflected in the languages used as a medium of instruction for educational systems, English being a predominant language. However, this educational model presents challenges for students in low-infrastructure and rural areas, where English proficiency tends to be limited.
We are currently in the process of conducting a randomized controlled trial where students are split into two language conditions: English and Leblango, which is a main local language used in Ugandan communities. In this RCT, we are investigating three main questions : 1) Perspectives on Language in STEM: views of students and facilitators on language use in STEM education 2) Language Impact on Learning: influence of instructional language on the student learning experience, STEM motivation, and career prospects 3)Technology's Role in Language Learning: impact of technology on learning effectiveness in English versus local languages.

We are looking for students who are passionate about conducting behavioral research in education and qualitative data analysis, with a strong focus on thematic analyses and in-depth exploration of qualitative data, including interview transcripts, discussion-based conversations, and other data sources.

Undergraduate Students, Masters Students, PhD Students

A defining dilemma of modern social media is helping users balance the social benefits they reap through online self-disclosure versus the privacy risks associated with those disclosures. Part of the challenge is that while the benefits of self-disclosure are often apparent, the harms are abstract and difficult to reason about which can lead to disclosures that people regret.

Privacy mirror is a chrome browser based plugin we are building with the intentions to empower users of SNS platforms through improving their awareness of self-disclosures in their posts that may pose a threat to their privacy. This project insights from HCI and privacy with NLP technology to highlight potentially harmful self-disclosures in users’ posts as they’re drafting them, and empower them in crafting alternative phrasings that offer more protection of their privacy while still allowing them to reap the benefits of online self-disclosures.

The student on this project will help front and backend development of the plugin itself. As part of this role, this independent study student will work to implement the UI and UX of the app based on existing wireframes, and help bring the plugin to a minimum viable product state. The student in this role may also have an opportunity to conduct evaluations of this intervention.

Undergraduate Students, Masters Students

We are developing an intelligent health assistant that leverages wearable devices for sensing and designs interventions aimed at delivering beneficial outcomes for those in need. Our evaluation process involves working with individuals diagnosed with dementia and ADHD. Currently, we are exploring several project ideas, including: (1) enhancing our sensing algorithms, (2) developing a user interface for clinicians to monitor patients, and (3) implementing a real-time intervention system designed to facilitate behavioral change.

Research Intern (Project Co-Leader)

Undergraduate Students, Masters Students, PhD Students

Bistable structures are interesting input devices in Human-Computer Interaction (HCI) due to their unique interactive properties. They can maintain the shape without continuous force input (e.g., our previous work: ConeAct), with precise and discrete configurations. Developing self-sensing materials based on bistable micro-structures is particularly exciting because it bridges the physical and digital worlds by translating tangible shapes into digital information. In this project, you will LEAD the development of fabrication methods for bistable sensing materials.

Task description:
- Fabrication Method Development: Develop fabrication methods for self-sensing bistable structures.
- Toolkit Development: Propose a toolkit for novice users that includes both software and hardware.
- Tangible Prototyping: Fabricate prototypes to demonstrate potential use cases.

Remote, On Campus

Undergraduate Students, Masters Students

Intelligent tutoring systems (ITSs) are online learn-by-doing activities with step-by-step coaching--a bit like a human tutor. They are very effective in helping students learn, so we'd like to broaden their use among teachers and students everywhere. How can we make them super easy for non-programmers, like curriculum developers or teachers, to create? Years ago, our lab invented CTAT--tools that enable authors to build their own ITSs without writing code. In order to improve access and ease of use, we've now rewritten CTAT as a web app using the Vue.js framework. We are looking for students who can help us improve and extend this browser-based version of CTAT. The work will develop skills in user-centered design, web application architecture, execution of user testing, and real-world application development.

https://github.com/CMUCTAT/CTAT/wiki/

Prototyping a Chatbot UI Supporting Agents Responding to a Complex Crisis Management Scenario

Research and Engineering, On Campus, Independent Study

We aim to build a chatbot agent that can monitor the environment in crisis management scenarios (e.g., ports after a tsunami). The agent is designed to interact with the users in a mixed-initiative manner and can help the users to do planning, retrieving and reviewing relevant information, allocate resources, and answer questions. We have an interdisciplinary team with researchers in Robotics Institute and HCII, building multiple modules supplying the chatbot, including simulators that mimic the physical environment, planners that help users craft possible next steps, etc.

We hope to find a research assistant working on frontend prototyping. While the actual task might vary as the project goes, the student will likely help design the interaction between the agent and the human, and develop the front end of the chatbot interface (using Figma).

Students with more advanced web development skills can also help integrate the agent with other parts (e.g., a display of an interactive map, a potential linked display to the working process of other components like simulators, and a hub monitoring the current resources like drones in the simulator). This will also require some reasonable understanding of the current message-sharing system, databases, and components we are building.

Remote, On Campus

Undergraduate Students, Masters Students

As work becomes more remote and distributed, a range of surveillance technologies are increasingly being used by businesses to track and monitor employees. While workers express concern about risks to their health, safety, and privacy, they also face a lack of transparency and autonomy around the use of these systems. This project aims to design and develop a public workplace surveillance tracker to create more transparency around this issue. We aim to populate this tracker using cases pulled from several data sources, including: news outlets, academic publications, legal cases, government agency reports, vendor websites, among others. Additionally, we will provide mechanisms for workers to submit their own cases with the ability to include personal testimonials of their experiences, while preserving their privacy. By offering detailed information from various credible sources and including firsthand accounts from workers, the tracker will serve as a resource to inform employees, researchers, and policymakers about current surveillance practices and trends in the workplace and help them to take action.

Research intern/ volunteer, Independent Study opportunity

We are designing and developing a teacher tool leveraging multimodal data analysis (intelligent tutoring systems, transaction log data, location sensors, possibly audio data). We are looking for students to help out with conducting this research project. Main research tasks that students are expected to help out with include primarily helping out user research & design, literature review, quantitative and qualitative data analysis. We are looking to conduct research study/ experiment with actual teachers and students users in the authentic classroom settings.

The student will take on both researcher and developer roles.

Undergraduate Students, Masters Students, PhD Students

Communities often rely on regulatory policies for governance. For instance, online communities like Reddit have content moderation rules that determine what can be posted in each subreddit. However, community members may interpret these rules differently or hold conflicting views on how they should be enforced, leading to inconsistent decision-making and frustration. In this project, we are developing an AI-assisted, end-to-end system that enables community members to collaboratively identify the root causes of their disagreements and iteratively revise their communal rules to resolve these conflicts. The student will build on our previous work to develop an end-to-end system and conduct user studies within real-world community contexts.

VTutor: An Open-Source SDK for Education with Generative AI-Powered Animated Pedagogical Agents

Remote, On Campus

Eason Chen, Jionghao Lin

Undergraduate Students, Masters Students, PhD Students

Textual feedback has long been central to education, but adding multimodal feedback—combining auditory and visual cues—can greatly enhance engagement and learning outcomes. Current LLMs are limited to text, leaving multimodal feedback unexplored in education. To address this, we launched the MuFIN project, focusing on Animated Pedagogical Agents (APAs) as virtual tutors – VTutor. Initial efforts using generative AI for feedback videos proved impractical due to high computational demands, with one-minute videos requiring two minutes of processing.

To overcome this, we developed VTutor using Unity, enabling real-time APA feedback with Unity WebGL and Large Language Models (LLMs). VTutor provides human-like feedback via voice and gestures, delivering an interactive, scalable learning experience accessible on any browser. A demo is available at https://vtutor.vercel.app (Note that when open demo, it may take some time to load).

Pilot studies on math and language tutoring showed VTutor improves engagement and retention. Its cost-effective iframe-based integration makes it ideal for K-12 and higher education, demonstrating potential for enhancing learning through generative AI and multimodal interaction.

Remote, On Campus

Mathematics education is foundational for advanced science and engineering disciplines, yet tutoring resources in higher education are often limited due to faculty and teaching assistant constraints. Large Language Models (LLMs), with their reasoning and language capabilities, offer scalable potential for automated math tutoring.

To explore this, I developed GPTutor Math, leveraging LLMs to acts as a Proof Assistant for students at college level math courses. The framework supports hinting, logical guidance, and solution critiques, tested in Carnegie Mellon University's Mathematical Foundations (15-151 / 21-127 / 21-128) courses with 315 students. Through human evaluations, interviews, and surveys, we analyze GPTutor's impact on comprehension and exam performance, comparing it to traditional university resources. This research highlights generative AI's role in advancing math education.

https://math.gptutor.tools/class/cm1tf3i690000sypn2qe7w81b/homework/cm1tf84xa0001sypnkcz1f4sa/cm1uz9z6g0002su34xurvell7?classId=cm1tf3i690000sypn2qe7w81b&homeworkProblemId=cm1tf84xa0001sypnkcz1f4sa&homeworkSubmissionId=cm1uz5e3w0001su34colail47

Undergraduate Students, Masters Students, PhD Students

We plan to develop innovative college-level math courses integrating Learning Sciences and Human-Computer Interaction advancements. Starting with Calculus. Welcome to join us!

Undergraduate Students, Masters Students

The goal of Comprehensive Personalized Programming Practice Environment (C3PE) is to create engaging learning experiences for computer science students. The system will recommend the most suitable learning activities for students based on their current knowledge level and offer personalized feedback to support their progress. To understand the kinds of practice opportunities that are most effective, we are analyzing data from past Intro Computer Science courses that used worked examples.

Exploring student engagement, motivation and interest with a multilingual Generative AI tutor in Africa

Undergraduate Students, Masters Students

This project investigates how students engage with a multilingual generative AI tutor. As an initial step, we analyze the chat logs between students and the AI tutor to understand the ways in which the students interact with AI in native languages, examining aspects such as language preference, sentiment, topics, communication styles.

Undergraduate Students, Masters Students, PhD Students

More detail about this role is available here:
https://docs.google.com/document/d/1ytk3AyxY_2d2VS4r1KRx56wqT189NikC44t25zFZHAw/edit?usp=sharing
Applicants MUST read the full listing.

Project Description
We are developing a game designed to teach self-regulation skills to middle school students. This position provides an excellent opportunity for students interested in game design, educational technology, and user research. As an Apprentice Game Designer, you will have the chance to contribute to the design process while gaining hands-on experience conducting field research and testing.
Summary of Work
As an Apprentice Game Designer, you will primarily focus on conducting field research at local schools, gathering feedback from middle school students, and running playtests. In addition, you will have opportunities to collaborate with the design team, contribute ideas, and assist in refining the game based on the data collected. This position is unpaid but can be taken as an independent study for course credit.

Undergraduate Students, Masters Students

We have multiple opportunities to work on projects in the areas of Augmented Reality, Virtual Reality, optimization-based user interfaces, audio interactions, multimodal interaction, AR/VR + applied ML, and others.

For example research, please see our website
https://augmented-perception.org/

Our projects typically require strong technical skills, particularly in at least one area of computer graphics, 3D programming, computer vision or applied machine learning. Furthermore, to ensure a successful project, we require a minimum time commitment of 8-10h / week.

If you are interested in pursuing one of the projects or have any questions, please send an email to Prof. Lindlbauer at davidlindlbauer@cmu.edu.

Include the following information:
- Current resume
- Experience with programming, designing, 3D creation
- How many hours a week can you dedicate (minimum 8-10h / week)
- Mode of project (Independent study, SURA,...)

Research Interview on Changes in Coding Practices and Learning with AI Assistants

Area: Social computing
Contact: Aoi Naito, Hirokazu Shirado

The widespread use of AI-powered assistants in software development has a significant impact not only on programmer’s productivity but also the whole programming community. Research on the digital trace data suggests that the information source for programmers is shifting from online knowledge sharing platforms such as Stack Overflow to AI-powered coding assistants represented by ChatGPT.
This independent research will assess the impact of AI on programmers’ subjective experience by interview method. Students will interview CMU Computer Science students and alums in software development roles to understand how AI-powered tools have changed their daily coding practices and the way to learn programming.

Analysis of digital trace data on the relationship between AI-use policies in GitHub repositories and community culture

Area: Social computing
Contact: Aoi Naito, Hirokazu Shirado

The widespread use of AI-powered assistants in software development has a significant impact not only on programmer’s productivity but also the whole programming community. Some open-source software projects have expressed concerns about licenses and intellectual property conflicts with AI-generated code, leading certain mature projects to regulate or even prohibit AI-powered tools.
This independent study will explore various GitHub public repositories and their policies regarding AI-powered coding assistants by analysing GitHub trace data. Also, we will perform comparative analysis on how different levels of AI adoption influence developers’ activities within repositories.

Remote, On Campus

Undergraduate Students, Masters Students, PhD Students

Mosaic is an Extensible Framework for Linking Databases and Interactive Views. As part of this project, you will develop new algorithms, interfaces, and evaluations for this framework used by hundreds of developers worldwide to create lightning fast machine learning and data science applications.

https://github.com/uwdata/mosaic/issues?q=sort%3Aupdated-desc+is%3Aissue+is%3Aopen+label%3Aprojects

AI Variety Cards: Supporting Responsible AI Ideation Across Various AI Model Types

Remote, On Campus

Undergraduate Students, Masters Students

Recent advances in Generative AI have introduced a diverse range of models into AI innovation, such as general-purpose foundation models, domain-specific models, customized models, and hybrid models. This diversity complicates the Responsible AI (RAI) innovation process, as AI product teams often face challenges in addressing feasibility, viability, desirability, and ethical issues, particularly during the early ideation stage. The goal of this project is to develop a novel Responsible AI toolkit – AI Variety Cards – to help AI product teams better understand and navigate the advantages, limitations, and ethical considerations of various AI models. This toolkit aims to facilitate more ethical decision-making by guiding AI product teams to select models that balance feasibility, technical viability, stakeholder desirability, and potential harms, contributing to our long-term research vision of developing and assessing innovative tools, processes, and methods that support the Responsible AI (RAI) ideation. By promoting ethical decision-making in AI model selection, our research will also help inform policies focused on transparency, fairness, and accountability in AI innovation, ensuring that AI systems are developed responsibly from the early stages.

Understanding the Implications of Identity, Culture, and Representation in Emerging Generative AI Systems

Adinawa Adjagbodjou, Lisa Egede

Undergraduate Students, Masters Students, PhD Students

In this work, we explore how user's identities, perceptions, and AI literacy shape their interactions with emerging AI technologies across various contexts. Through human-centered methods, this project seeks to explore implications of the design of these systems by further understanding user's experiences and reflection on navigating them.

Undergraduate Students, Masters Students

We aim to conduct a mix-method study to understand users' perceptions on LLM safety guardrails. The main research questions will be around how everyday users make sense of the "invisible" safety guardrails implemented within commercial LLM and how these safety guardrails are impacting the end users.

Enhancing Personalization and Accessibility in Virtual Reality Experiences

VR Development and Prototyping

Undergraduate Students, Masters Students

This project explores how personalization and accessibility can enhance user experiences in virtual reality (VR). By focusing on inclusive design, it aims to develop adaptable VR interactions that empower users of all abilities and foster positive engagement in immersive environments.

Co-Create Your Own Adventure: Collaborative Fanfiction for Mental Wellness in LGBTQ+ Populations

Remote, On Campus

Undergraduate Students, Masters Students

The goal of this project is to create an online collaborative fanfiction writing tool for mental wellness within the LGBTQ+ community. This tool is for novice writers to share stories and gain social support from trained near-peer mentors. This semester’s focus will be on 1) designing and testing a fanfiction collaborative writing interaction, 2) building focus groups with local LGBTQ+ mental health organizations, and 3) creating a training guide for near-peer mentors. Applicants from all research experience levels are welcome!

Enhancing Children’s Reading Engagement and Comprehension through AI-powered Digital Companions

Remote, On Campus

Undergraduate Students, Masters Students

Project Overview: This project aims to develop an innovative AI-powered digital companion to enhance children's reading experiences, specifically for ages 10-12. This digital companion will act as an interactive "reading buddy," engaging children in reading activities, enhancing their comprehension, and supporting their literacy skills through personalized interactions. This research will contribute to Human-Computer Interaction (HCI), educational technology, and AI, focusing on creating supportive, engaging learning environments.

Remote, On Campus

Undergraduate Students, Masters Students

With the competency of text-to-image generative models rapidly increasing, clinical machine learning researchers are making significant advances in developing models capable of generating complex medical imagery from text prompts. Despite numerous generative text-to-image models being developed for radiology, research has overlooked how stakeholders would benefit from and use GenAI in radiology work. Without stakeholders’ input, we risk the potential of building GenAI tools that are not useful or appropriate for their work. Using our GenAI-based user interface powered by a text-to-3D CT scan generative model as a design probe, our work aims to shape the use of GenAI in radiology work and education.

Independent Study Research Assistant

Remote, On Campus

Undergraduate Students, Masters Students

Clinical decision support systems provide AI-driven insights that can help clinicians treat patients. But it is hard to use these systems' current interfaces and the ways they communicate information. We are conducting multiple studies with physicians in intensive care, including design interviews, prototyping, and a large-scale survey, to understand how they might use new interfaces for decision support. We are looking for 1-2 self-directed students with an eye for design and attention to detail, to help us create the prototype decision support systems and run the studies. Students will gain exposure to designing for healthcare, as well as practice using HCI methods such as wireframe prototyping, thematic analysis, and survey design. If interested, please email venkats@cmu.edu with a resume, why you want to work on the project, and a description of a prior HCI project you did and how you analyzed the results.

Lumilo: Augmented reality for teachers in the smart classroom of the future

Remote, On Campus

Undergraduate Students, Masters Students

The mission of the CATS lab is prototyping new technologies for the smart classroom. One of our projects deals with developing smart glasses for teachers, where they can see, above each student's head, how well each of them is working/progressing with math learning software - while the students are working on the software.

Mixed Reality Tools to Support Teacher Analytics and Decision-Making

Remote, On Campus

Independent Study, Research Volunteer

Undergraduate Students, Masters Students, PhD Students

Are you passionate about developing impactful educational technology? In collaboration with real edu tech company and real classroom, this project focuses on designing and building mixed-reality smart glasses (using Xreal) to enhance K-12 education. These glasses will integrate real-time analytics from Intelligent Tutoring Systems (ITSs), providing teachers with actionable insights and recommendations to guide their decision-making during the class. You will use Unity3D as the development platform to create an ITS with AR technology, with the exciting opportunity to test it in a real classroom environment!

https://drive.google.com/file/d/1eZqi0TRKGHiMJMUakRc9ttdzvxCzG1Ig/view

AI-Driven Teacher Support: Building a Recommendation System for Classroom Decision-Making

Remote, On Campus

Independent Study, Research Volunteer

(Position has already been filled)

In a busy classroom, teachers often struggle to allocate their attention effectively to make optimal decisions about whom to assist and how to provide support. This project leverages data collected from real classrooms to address this challenge. Your role will involve developing a machine learning-based recommendation system that uses students' activity logs from an Intelligent Tutoring System to guide teachers in making informed and timely decisions to support student learning.

Designing Tangible Interfaces and AI Collaborators to Enhance Children's AI Literacy

HCI researcher (child-computer interaction)

Remote, On Campus

Independent Study, Research Volunteer

Undergraduate Students, Masters Students, PhD Students

How can AI glasses (e.g., Ray-Ban) empower children to learn programming and AI literacy through tangible interaction? In this project, you will design a tangible interface to support programming or AI literacy and explore how AI, as a collaborative partner, can help children grasp key concepts and keep them engaged in learning.

Remote, On Campus

Undergraduate Students, Masters Students, PhD Students

We are seeking a motivated student with experience in R and/or Python to support research on learning in intelligent tutoring systems. The primary focus of this role is analyzing log data to compare student learning outcomes across different intelligent tutoring systems. This involves applying data processing, statistical analysis, and machine learning techniques to identify patterns in student interactions and learning gains.

In addition to data analysis, the role includes supporting the preparation of study materials for upcoming classroom experiments and assisting with recruitment efforts for a Fall study. This position provides a unique opportunity to contribute to cutting-edge research on technology-enhanced learning while working with an interdisciplinary research team.

If you are interested in applying computational methods to education and gaining experience in data-driven learning science research, we encourage you to apply!

Undergraduate Students, Masters Students

Communities often rely on regulatory policies for governance. For instance, online communities like Reddit have content moderation rules that determine what can be posted in each subreddit. However, community members may interpret these rules differently or hold conflicting views on how they should be enforced, leading to inconsistent decision-making and frustration. In this project, we are developing an AI-assisted, end-to-end system that enables community members to collaboratively identify the root causes of their disagreements and iteratively revise their communal rules to resolve these conflicts. The student will build on our previous work to develop an end-to-end system and conduct user studies within real-world community contexts.

Undergraduate Students, Masters Students

The goal of this project is to leverage Unity and its spatial audio capabilities to develop a plugin for creating Augmented Reality and Virtual Reality programs that use spatial audio. The plugin will be used for running studies (example “how good can users localized sound in virtual spaces?”), as well as for interactive demonstrators.

This project is a collaboration between Prof. David Lindlbauer and Prof. Laurie Heller (https://www.auditorylab.org/). If you are interested in pursuing this project or have any questions, please send an email to Prof. Lindlbauer (davidlindlbauer@cmu.edu) and Prof. Heller (hellerl@andrew.cmu.edu).

Include the following information:
- Current resume
- Experience with programming, designing, 3D creation
- How many hours a week can you dedicate (minimum 8-10h / week)
- Mode of project (Independent study, SURA,...)

An Informal Learning Management System embedded Narrative Adventure Game for Primary Education

Remote, On Campus

pmd.ferreira@campus.fct.unl.pt

Undergraduate Students, Masters Students

This project regards the prototyping of an Informal Learning Management System embedded narrative adventure game for primary school children and their families. The game is to introduce a parallel fantasy universe and to link children’s everyday learning activities to their character’s development and story progression in this universe. It aims to support families and to motivate children’s participation in the management of their experiences out-of-school.

AIML developer / Web developer

Undergraduate Students, Masters Students, PhD Students

We are developing an auto-documentation tool to support medical teams in the ICU. In our initial study, we explored the challenges and opportunities associated with computing systems in ICU settings. Additionally, students will have opportunities to participate in subsequent studies, co-author research papers, and collaborate closely with Prof. John Zimmerman, Prof. Jodi Forlizzi, and our partners at other institutions (including doctors and medical teams from UPMC).

Undergraduate Students, Masters Students

We are developing tools to help designers ideate using LLMs. By doing so, we aim to unlock new ways of combining human and LLM intelligence, leading to better decision-making and enhanced creativity. A recent example of our work is InkSpire (https://arxiv.org/abs/2501.18588).

Our ongoing work focuses on augmenting creativity by helping users explore a broad range of inspirations while also diving deep into specific areas of interest. The tool is designed to support users in effectively analyzing idea trade-offs and avoiding fixation. We have some initial designs and code and are looking for two programmers to join our team. The primary responsibilities will include refining the design, developing the front end based on our current prototype, implementing system interactions, and visualizing data.

Proficiency in front-end development (our current prototype is built with React);
Proficient in prompt engineering;
Background in NLP or design, along with experience in data visualization (preferred)

Research Assistants, Collaborators

Remote, On Campus

This research investigates how AI researchers reason about benefits in their work. Specifically, we’re interested in the tension between work that produces immediate benefits and work that makes a technical advance with only a promise that someone else might later produce a societal benefit. We aim to explore what they think counts as AI research, what constitutes research excellence, and who they see as the users of and/or beneficiaries of their research.

Detecting Collaborative Moments in Computer-Supported Collaborative Learning

Remote, On Campus

haokanga@andrew.cmu.edu, ckwon2@andrew.cmu.edu

Undergraduate Students, Masters Students

We are looking for students interested in these topics: computer-supported collaborative learning, technology-enhanced learning, Large Language Models (LLMs), and learning sciences.

This project investigates how students collaborate in computer-supported collaborative learning (CSCL) environments. We are also exploring the use of Large Language Models (LLMs) to automatically annotate collaborative moments, paving the way for scalable, AI-assisted analysis of teamwork in CSCL.

We are looking for motivated undergraduate or graduate students who: 1) are interested in research in CS education 2) have basic programming skills in data science (e.g., Python, Jupyter, pandas, etc.) 3) are curious about how AI tools like LLMs can support educational research!

Remote, On Campus

Undergraduate Students, Masters Students

Summer Research Assistant Position: Designing Scenarios for AI Evaluation
We are seeking a paid research assistant to join an interdisciplinary project focused on designing better ways to evaluate the risks and benefits of AI systems. As AI technologies are increasingly deployed in high-stakes domains, organizations need clearer, more robust ways to test these systems in context. This project aims to develop processes and tools for designing evaluation scenarios—structured exercises that simulate real-world use and uncover system strengths, vulnerabilities, and unintended consequences.

The work will span early-stage design and analysis and may include:
Assisting with large-scale data analysis or tool-building to automate parts of the scenario design process (Python, statistics, and/or NLP experience is a plus)
Helping draft research publications, with potential co-authorship opportunities
Supporting qualitative research (e.g., analyzing interviews with industry practitioners to identify evaluation needs)
Conducting literature reviews on AI evaluation, risk assessment, and human-AI interaction


Required Qualifications:
Experience with Python and basic statistical analysis
Interest in research in AI ethics, evaluation, and human-centered computing
Strong communication skills and attention to detail

Preferred Qualifications (not required):
Prior course in statistical analysis, machine learning, natural language processing, and/or human-computer interaction
Demonstration of strong academic performance (e.g., strong major GPA)
Previous research experience in human-computer interaction, AI/ML, or a related field

In this project you will work directly with Alice Qian Zhang for day-to-day scenario development. You may also meet with Hong Shen, Laura Dabbish, and other researchers as part of our broader multidisciplinary team. This will be a paid hourly role for Summer 2025. The time commitment expectation is at most 20 hours/week from June through August 2025 with some flexibility on exact dates. Work modality is flexible for either in-person or remote.

How to apply:
Please prepare a cv/resume that includes major GPA along with no more than 500 words (roughly 2-3 paragraphs) explaining (1) what specific aspects of the project you are interested in, (2) what aspects of the project you are well-positioned to contribute to, and (3) how participation will help your broader interests and goals. Email the above information to <aqzhang@andrew.cmu.edu> with subject header “RA Position Interest - [insert your name].” Applications will be reviewed on a rolling basis.

Required Qualifications:
Experience with Python and basic statistical analysis
Interest in research in AI ethics, evaluation, and human-centered computing
Strong communication skills and attention to detail

Preferred Qualifications (not required):
Prior course in statistical analysis, machine learning, natural language processing, and/or human-computer interaction
Demonstration of strong academic performance (e.g., strong major GPA)
Previous research experience in human-computer interaction, AI/ML, or a related field

Remote, On Campus

Undergraduate Students, Masters Students

Intelligent tutoring systems (ITSs) are online learn-by-doing activities with step-by-step coaching--a bit like a human tutor. They are very effective in helping students learn, so we'd like to broaden their use among teachers and students everywhere. How can we make them super easy for non-programmers, like curriculum developers or teachers, to create? Years ago, our lab invented CTAT--tools that enable authors to build their own ITSs without writing code. In order to improve access and ease of use, we've now rewritten CTAT as a web app using the Vue.js framework. We are looking for students who can help us improve and extend this browser-based version of CTAT. The work will develop skills in user-centered design, web application architecture, execution of user testing, and real-world application development.

JavaScript, HTML. Experience with Vue.js or React is helpful but not required.

https://github.com/CMUCTAT/CTAT/wiki/

Undergraduate Students, Masters Students

This independent study explores how end users intuitively express privacy and access control preferences through freehand drawings. The project will conduct a crowdsourced study to collect a dataset of hand-drawn privacy policies, analyze emerging patterns, and develop a visual language for privacy specification. The goal is to create a system analogous to Vega-Lite’s role in data visualization—a concise, expressive visual grammar, but for privacy and access control policies instead of general data.

Interest in privacy, HCI, and machine learning.

Experience with data collection and crowdsourced studies (e.g., Prolific).

Familiarity with machine learning and model fine-tuning (e.g., PyTorch, Hugging Face).