Interested in research opportunities for course credit
Email lsec@iu.edu to speak with the Luddy Office of Student Engagement and Success.
Research Experiences for Undergraduates (REU) opportunities at the Luddy School of Informatics, Computers, and Engineering at IU Bloomington connects undergraduate students from IU and other institutions with faculty research opportunities.
REU participant requirements:
U.S. citizen, U.S. national, or U.S. permanent resident.
An undergraduate student.
REU participants must also:
Complete Responsible Conduct of Research (RCR) training through CITI at IU.
Participate in a Poster Event at IU/Luddy.
REU Research Opportunities at Luddy IUB
Learn more about the available academic and summer REU research opportunities and how to apply for the positions in the table below. Summer REU opportunities are open to non-IU students also.
The RAD-IUB REU Program is an eight-week research internship experience running from June 2 - July 18 in Bloomington, IN. Students work closely with their mentors on cutting-edge projects in Microelectronics Design, Space Systems, Radiation Effects, Electronics, and other exciting research areas. https://create.iu.edu/reu.html
Required:
U.S. Citizen
Currently enrolled in a College or University Physics, Engineering, Mathematics, or Computer Science Program.
For this project, we have point cloud data of a state park collected after storm that shows standing trees, fallen trees, root ball and pits formed after tree throw. One goal of this project is to develop a visualization tool that aligns bird-eye view (BEV) images and point clouds, and to facilitate the annotation of different parts of the trees, root balls and pits from different views. The annotated data will be used to train different point-cloud object detection and segmentation models to segment out different parts of the trees (fallen or standing), root balls, and pits. In this project, the students will learn how to work on the visualization tool, and use it to annotate point cloud data, as well as training and evaluating deep learning models for 3D point clouds (combined with BEV data to improve performance).
Required:
Familiar with Python and can use Python to writing small programs by calling various libraries.
Being careful and hold a high standard of their work. This is important since the quality of our AI models and depends on the quality of data annotation.
Eagerness to learning the deep learning pipeline including data preparation, model training with a GPU, and model evaluation.
Preferred:
Familiar with PyTorch and have some experience in deep learning projects.
Familiar with JavaScript and WebGL
Apply by submitting the online application and attach the following documents:
Resume
Two letters of reference from instructors or research mentors.
The student will contribute to the curation of a satellite image dataset for disaster events such as flooding and hurricanes. We have already downloaded a large number of post-disaster satellite images from Maxar, and the first task of the student is to identify regions that are flooded (rare events) to increase their presence and importance during the pre-training of geo-foundation models. We have designed a few geo-foundation models that take digital elevation model (DEM) data as additional terrain context, and the student will experience the process of distributed training of geo-foundational models. For fine-tuning for the flood mapping application, student will help annotate the flooded areas in the identified regions using our active learning tool, and then finetune our geo-foundational models on the annotated dataset using terrain-aware loss function.
Required:
Familiar with Python and can use Python to writing small programs by calling various libraries.
Being careful and hold a high standard of their work. This is important since the quality of our geo-foundation models and fine-tuned flood mapping models depends on the quality of data curation and annotation.
Eagerness to learning the deep learning pipeline including data preparation, model training with a GPU, distributed model training with many GPUs, and model evaluation.
Preferred:
Familiar with PyTorch and have some experience in deep learning projects.
Knowledge about distributed model training is a plus but not a must, as we can teaching how to do that.
Apply by submitting the online application and attach the following documents:
Resume
Two letters of reference from instructors or research mentors.
The Health Insurance Portability and Accountability Act (HIPAA) mandates strict guidelines to protect the privacy and security of sensitive patient health information (PII), requiring the removal of personally identifiable information (PII) from medical records before they can be used for research. This REU project addresses the need for HIPAA-compliant de-identification by developing an automated system that leverages a lightweight Large Language Model (LLM) to identify and mask PII in clinical documents while preserving the data's research utility. The de-identification process will be securely deployed within a confidential computing environment using AMD's Secure Encrypted Virtualization (SEV) technology. This approach ensures maximum data privacy and security, offering students hands-on experience at the intersection of natural language processing, healthcare data privacy, and secure computing, and tackling critical challenges in protecting patient information in research and clinical applications.
Required:
None required.
Preferred:
Any programming language is helpful, but not required.
Apply by submitting the online application and attach the following documents:
Resume
Letters of recommendation recommended, but not required.
This REU project will explore potential privacy leaks in the vectorized representation of patient health records and investigate secure methods to mitigate these risks within confidential computing environments. With the rise of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) techniques, healthcare researchers are increasingly utilizing vectorized patient data to enhance information retrieval and personalized medicine applications. While these technologies offer tremendous potential to advance healthcare research, they also introduce privacy concerns, as vector embeddings can inadvertently reveal sensitive information. This project will involve assessing privacy vulnerabilities in health data embeddings and developing confidential computing solutions, such as AMD’s Secure Encrypted Virtualization (SEV), to secure LLM/RAG applications using vector databases. Participants will gain hands-on experience in applying privacy-preserving strategies in AI-driven healthcare research, contributing to a critical area of health informatics and data security.
Required:
None required.
Preferred:
Any programming language is helpful, but not required.
Apply by submitting the online application and attach the following documents:
Resume
Letters of recommendation recommended, but not required.
REU participants receive a stipend determined by the faculty/PI. REU stipends and support are processed each semester as scholarship. It is strongly recommended that, prior to accepting an REU position at IU or another institution, students contact a Financial Aid Advisor through Student Services to discuss the impact of the stipend and support on their current and future financial aid packages.*
*IU students involved in REU programs at other institutions are required to notify IU OSFA of all scholarships received as stipend and support. Similarly, students from other institutions are required to notify their home institution of the stipend and support received as an REU at IU.
Doing an REU showed me that computing work doesn't always have to be coding.
