The 10 currently listed mentors are for the Fall 2025 semester.
If you have any questions about the information below, please reach out to the URSA Committee at URSA-group@office365.illinois.edu.
Why become an URSA Mentor?
| Mentor | Description |
|---|---|
Aditi Tiwari (Computer Science) | Aditi Tiwari is a second-year Ph.D. student in Computer Science at the University of Illinois Urbana-Champaign, co-advised by Professors Heng Ji and Klara Nahrstedt. She also holds an M.S. in Computer Science from UIUC. Her research focuses on building multimodal, protocol-driven agents that can perceive, reason, and act in complex visual environments by composing modular tools. She investigates how intelligent agents can learn grounded visual reasoning and control through interaction, using frameworks such as the Model Context Protocol (MCP) to orchestrate tools for video understanding, image editing, and dynamic scene interpretation. Her broader interests include vision-language modeling, embodied world models, and interactive perception for real-world, resource-constrained AI systems. Description of Possible Projects: The student will work on a project at the intersection of vision, language, and tool-based reasoning. The goal is to build a multimodal agent for instructional video understanding and transformation. Given a narrated video (e.g., "how to change a tire"), the agent will learn to segment relevant steps, identify key objects and actions, and generate guided edits—such as highlighting tools in use, removing distractions, or assembling concise summaries. Project Scope (Fall 2025 timeline): Weeks 1–3: Familiarize with basic models (e.g., CLIP, SAM, Whisper), select a small set of instructional videos, and define edit goals. Weeks 4–7: Build an initial pipeline for tool-conditioned video segmentation using narration cues and object tracking. Weeks 8–11: Implement grounded planning—using simple agent logic or protocol-based calls (MCP) to apply edits based on video semantics. Weeks 12–14: Evaluate outputs using sample metrics (e.g., clarity, consistency), and prepare a short write-up or demo presentation. Deliverables: Functional video segmentation + editing prototype Annotated dataset of prompt-edit pairs with reasoning Final demo or short research write-up Desired Skills and Time Commitment: Required: Strong Python skills; willingness to explore visual and language models Preferred: Some background in PyTorch, video processing (OpenCV or ffmpeg), or experience with segmentation tools (e.g., CLIP, SAM) Estimated Weekly Commitment: ~8–10 hours/week Meeting Structure: Weekly one-on-one check-ins and occasional group meetings with other mentees |
| Erin Nicholas (Chemistry)  | Erin Nicholas is a rising fourth-year Chemistry graduate student in the Vura-Weis group. She received a B.S. in Chemistry and a B.S. in Mathematics from Arkansas State University. At Arkansas State, she was introduced to laser spectroscopy and became entranced with these techniques. At Illinois, her research utilizes extreme ultraviolet spectroscopy (XUV) to study how small changes in molecules affect their spectra. She is also working on improving signal-to-noise in XUV measurements. Description of Possible Projects: Project 1: Investigating how small changes in iridium or platinum-based molecules will affect spectra. Mentees will assist in experiments for investigating how small changes in iridium or platinum-based molecules will affect spectra. We will create approximately 100 nm thick thin-films of iridium complexes, perform extreme ultraviolet spectroscopy (XUV) on these films, analyze the collected spectra using statistics, and compare the spectra to information from simulations of these molecules. This research area will help us answer questions about how metals such as iridium and platinum behave in X-ray absorption spectroscopy. Preferred Majors: Chemistry, Chemical Engineering, Materials Science & Engineering, Electrical Engineering, Nuclear Plasma and Radiological Engineering, Physics Preferred Skills: Knowledge of general chemistry concepts and desire to perform experiments safely. Required Courses: Either CHEM 102 or 202. Optional Courses: CHEM 104/204 Timeline: First week will be dedicated to acclimating to lab, completing DRS safety trainings and in-person safety trainings. Nine weeks will be dedicated to sample preparation, data collection, and literature readings. Final week will be dedicated to poster creation and practice presentation with the Vura-Weis group prior to the symposium. Deliverables: XUV spectra and poster for symposium. Project 2: Data mining to create a machine learning algorithm for our extreme ultraviolet spectra. Mentees will assist in the creation of a machine learning algorithm to aid in our analysis of the XUV spectra. In order to get the algorithm running, mentees will be performing data mining and looking through lab notebooks to ensure we have a large enough database to help create our machine learning algorithm. This will most likely not have any lab-time; however, you will still undergo the necessary safety trainings in the off-chance you need to be in the lab. If the student has a strong desire to do this project and be in lab, please let me know and we can discuss how to amend the project to do so. Preferred Majors: Chemistry, Chemical Engineering, Computer Engineering, Materials Science and Engineering, Physics Preferred Skills: Knowledge of python, knowledge of general chemistry concepts, desire to learn. Required Courses: Either CHEM 102 or 202. Optional Courses: CS 101, CS 173 Timeline: First week will be dedicated to any necessary safety trainings and acclimating to being a part of the group. Nine weeks will be dedicated to the data mining process. Final week will be dedicated to poster creation and practice presentation with the Vura-Weis group prior to the symposium. Deliverables: Poster for symposium. Desired Skills and Time Commitment: Project 1: Basic understanding of introductory chemistry concepts, ones learned in general chemistry courses. Desire to learn and be safe while in lab. The laboratory experience includes both a wet lab and a laser lab, so ensuring safety of yourself and those around you is a major part of this experience. The time commitment is 5-10 hours/week, but since experiments are rather involved, communication on available times will be necessary to coordinate when you can join and help run experiments. Python knowledge can help, but not required. Project 2: Python knowledge. Basic understanding of introductory chemistry concepts. Desire to learn. Time commitment is 5-10 hours/week. We will meet twice per week, so we will need to coordinate what times work well for both of us. |
| Javier Balta (Mechanical Science and Engineering)  | Javier Balta is a graduate researcher in robotics and materials at the University of Illinois Urbana-Champaign, with a background in microengineering from EPFL. His work focuses on additive manufacturing technologies for space applications, combining hands-on experimentation with system-level design. He is currently developing an automated materials testing system, drawing on years of personal experience building microcontroller-based projects and custom hardware. He is excited to mentor motivated students and share what research can look like across disciplines—from concept to lab reality. The robotic platform at the center of this work is designed to accelerate the discovery and testing of frontally polymerizable polymers. It features a robotic arm that autonomously mixes monomers, powders, and other liquid components, then runs a series of experiments to evaluate material performance. These include front speed measurements using custom-designed molds, differential scanning calorimetry (DSC), and proto-rheology tests such as vial tilting to assess viscosity. The two proposed student projects aim to refine and stabilize the existing setup while helping implement subsystems for the DSC and proto-rheology experiments that are still under development. Description of Possible Projects: Project Title 1: Automation and Prototyping for High-Throughput Polymer Testing This project involves contributing to the development of a robotic platform for high-throughput polymer formulation and characterization. The student will focus on automating one or more subsystems, which may include designing and 3D-printing mechanical components, wiring sensors and actuators, programming microcontrollers (Arduino), and verifying the system both as a standalone module and within the larger integrated setup. Attention to detail and a tidy work ethic in the lab are essential, as this work often involves wet lab procedures alongside electronics. Recommended background: Experience (or strong interest) in Arduino or embedded systems programming, CAD (e.g. Fusion 360 or SolidWorks), and hands-on prototyping. Familiarity with lab work and safe handling of materials is a plus, but not required. Majors or minors in Mechanical Engineering, Electrical Engineering, Robotics, Materials Science, or similar fields would benefit most, but all motivated students are welcome. Project Title 2: Experiment Workflow Design and Automation Logic This project will support the experimental pipeline of the same high-throughput platform, with a focus on refining experiment workflow and contributing to the scheduling logic of experiments. The student will assist with daily experiment execution, sample preparation, and documentation. A key component will be developing or improving the Python-based logic that governs the automation sequence—deciding when and how different parts of the system should interact to run efficient experiments. Recommended background: Some Python experience or willingness to learn quickly is necessary. Comfort working in a wet lab environment, following safety and preparation protocols, is also important. Majors or minors in Computer Science, Chemistry, Materials Science, or related disciplines are especially relevant. General Structure & Timeline Both projects will begin with a ~4-week onboarding and training period, during which the student will become familiar with the lab environment, tools, and ongoing projects. During this time, they’ll also be encouraged to address any skill gaps through self-paced learning (with guidance). Once familiar, students will identify their specific contribution focus, and we will collaboratively set semester goals. The remaining semester will be divided into three work phases, each ending in an informal review to assess progress and adjust goals if needed. Final deliverables will include a short presentation or report summarizing their work, organized notes, and ideally, a working prototype or demonstrable improvement to the system. Desired Skills and Time Commitment: No specific prior skills are required for these projects, though any experience with programming (especially Arduino or Python), 3D modeling, electronics, or hands-on lab work—whether from classes or personal hobbies—will definitely be useful. That said, what's most important is a strong willingness to learn quickly, both during lab hours and independently when needed. These projects are designed to give students their first exposure to real, applied research, so while I don’t expect mastery, I do expect curiosity, initiative, and a readiness to build new skills as we go. Students who enjoy tinkering, building, or solving practical problems will feel right at home, especially since the work is closely aligned with the kind of creative, hands-on experimentation often found in hobbyist projects. The lab is well-equipped, and I’m actively working on the same systems they'll be contributing to, so students will get guidance and mentorship throughout—but with room to propose ideas and take ownership of parts of the work. I’m very flexible with scheduling, but I prefer students to come in for 2–3 hour blocks at a time, since that allows us to take full advantage of the equipment and workflow setup. There’s no hard prerequisite in terms of coursework; any class that brings relevant technical or analytical skills (intro to programming, engineering fundamentals, basic chemistry or physics, etc.) can be helpful, but isn't required. |
Heng-Sheng Chang (Mechanical Science and Engeineering) | Heng-Sheng Chang is a postdoctoral researcher with a passion for advanced science, technology, and innovation. He earned his B.S. in Mechanical Engineering from National Taiwan University in 2017 and completed his M.S. and Ph.D. in 2025 at the University of Illinois Urbana-Champaign (UIUC) as a member of the Mehta research group. His doctoral work focused on modeling, control, and estimation of soft actuators, advancing the field of soft robotics through design, simulation, and experimental evaluation. His current research involves time series signal estimation, prediction, and generation. Description of Possible Projects: A Mathematical Framework for Language Models This project is a deep dive into the theoretical underpinnings of modern language models, specifically focusing on how they can be understood through a new mathematical framework. Mentees will help explore the "dual filter," an algorithm that models causal nonlinear prediction in a way that parallels the architecture of decoder-only transformers. This research is significant because it provides a theoretical link between optimal control and filtering, solving a long-standing problem in control theory. Students will get hands-on experience by conducting numerical experiments using parameters that mimic real-world transformer models to illustrate the algorithm's performance. Timeline: - Weeks 1-4: Introduction to the concepts of Hidden Markov Models (HMMs), nonlinear filtering, and the core principles of optimal control as they relate to prediction. Introduction to cloud computing. - Weeks 5-8: Experimenting with the filter algorithm using Python. Running and tuning numerical experiments to test the algorithm's predictive capabilities with realistic parameters. - Weeks 9-12: Analyzing and interpreting the results of the experiments to understand how the model captures correlations and generates predictions. Finalizing a project report and a presentation on the findings. Desired Skills and Time Commitment: Proficiency in Python, along with a foundational understanding of linear algebra |
| Michael Harrigan (Aerospace Engineering)  | Michael is a first-year PhD student in Aerospace Engineering working with the Laboratory for Advanced Space Systems at Illinois (LASSI) and Dr. Rovey on dual-mode monopropellant small satellite propulsion systems as well as other space systems and payloads. His primary interests are in space systems, program management, and systems engineering. His recent work involves developing a dual-mode monopropellant propulsion feed system for a suborbital test flight and requirement generation for a dual mode propulsion satellite. Description of Possible Projects: Project 1: FreeFlyer Script Library for Mission Analysis Tasks Objective: Develop a library of modular FreeFlyer scripts that automate common mission analysis tasks. Scripts will cover essentials like orbit propagation, ground pass visualization, delta‑v calculations, and perturbation toggling. The project emphasizes clean, reusable code with documentation so other researchers can easily integrate snippets into larger simulations. Majors: Aerospace Engineering, Computer Engineering, Physics, or related fields with interest in mission design and scripting. Skills: Basic coding proficiency (Python, MATLAB, or similar), willingness to learn FreeFlyer, and interest in spacecraft mission analysis and software tool development. Weeks 1-2: Install FreeFlyer, learn the interface, and review any existing scripts to understand structure and identify areas for new development. Weeks 3-4: Draft initial script templates, focusing on simple, broadly useful functions, and define a consistent structure and naming convention for the library. Weeks 5-6: Continue adding scripts covering a range of tasks, testing each one for functionality and consistency. Weeks 7-8: Add comments, inline documentation, and usage notes to each script, and gather feedback from graduate students and other lab users on what would make the scripts easier to use. Weeks 9-10: Refine the scripts for flexibility and reliability by improving inputs, adding error handling, and cleaning up the organization of the code. Week 11: Write a short user guide that explains how to access, run, and adapt each script in the library. Week 12: Present the completed library to the lab and hand off the repository for future expansion. Project 2: Satellite Design Support for MonARCH Objective: Support ongoing LASSI CubeSat missions by performing design tasks for specific subsystems. Work may include CAD model updates, component trade studies, interface diagrams, and basic performance analyses (e.g., power generation or mass budgets). The project focuses on connecting individual subsystem tasks to overall satellite architecture while producing documented outputs for reviews and integration. Majors: Aerospace Engineering, Mechanical Engineering, Electrical and Computer Engineering, or Systems Engineering. Skills: Introductory knowledge of satellite systems, experience with CAD or analysis tools (e.g., NX, MATLAB), and an interest in systems-level design work. Weeks 1-2: Get familiar with current LASSI missions and identify specific design support needs, which could include electrical diagrams, parts selection, or subsystem documentation. Weeks 3-4: Gather reference data, datasheets, and prior designs, then start on foundational tasks such as drawing updated wiring diagrams or listing interface points between subsystems. Weeks 5-6: Begin a focused study or design task such as selecting components for a power distribution board, mapping signal paths, or drafting a block diagram for a subsystem. Weeks 7-8: Review progress with the lab, incorporate feedback, and refine diagrams, tables, or parts selections based on comments and updated mission needs. Weeks 9-10: Expand the work into a clear, consistent package by cross‑checking documentation, labeling all diagrams, and filling any gaps in the design record. Week 11: Compile deliverables, such as wiring diagrams, component lists, or subsystem notes, into an organized, shareable format. Week 12: Present the work at a lab meeting, transfer files to the team, and recommend logical next steps for future students to continue the effort. |
Nusrat Chowdhury (Mechanical Science and Engeineering)  | Nusrat Chowdhury is a Ph.D. candidate in Mechanical Science and Engineering at the University of Illinois Urbana-Champaign, where she researches thermal transport and polymer engineering under Professor David Cahill. Her work focuses on developing next-generation polymeric materials with tailored thermal properties for energy and electronics applications. She has published her research in leading journals, including ACS Applied Polymer Materials, and presented at national and international conferences. Beyond research, Nusrat is passionate about mentoring students and supporting women in STEM. She has guided undergraduate researchers, led STEM outreach initiatives, and enjoys helping aspiring engineers navigate research and career development. Description of Possible Projects: Project 1: Synthesis and Characterization of Novel Polymers Description: This project involves synthesizing new polymeric materials and characterizing their thermal and mechanical properties. Students will gain hands-on experience with polymer synthesis, thermal analysis (TGA, DSC), and mechanical testing. The goal is to establish how polymer structure influences its performance in advanced material applications. Desired Background/Skills (Optional but Preferred): Major/Minor: Materials Science, Chemistry, Chemical Engineering, or Mechanical Engineering Skills: Basic chemistry lab experience, interest in material characterization, attention to lab safety Timeline & Deliverables: Weeks 1–2: Safety training, literature review, and learning synthesis protocols Weeks 3–8: Polymer synthesis and sample preparation for testing Weeks 9–12: Thermal and mechanical property measurements (TGA, DSC, tensile testing) Final Deliverables: A short report summarizing synthesis procedures and results Presentation of property–structure relationships to the research team Project 2: Machine Learning for Predicting Polymer Properties Description: This project focuses on developing a machine learning model to predict thermal conductivity and mechanical properties of polymers using an existing experimental database. Students will collect, clean, and analyze data, then implement and evaluate models to improve predictive accuracy. Desired Background/Skills (Optional but Preferred): Major/Minor: Computer Science, Data Science, Mechanical or Materials Engineering with coding experience Skills: Python or R, experience with machine learning libraries (scikit-learn, TensorFlow, or PyTorch), basic understanding of data analysis Timeline & Deliverables: Weeks 1–3: Data collection, cleaning, and exploratory analysis Weeks 4–8: Machine learning model development and testing Weeks 9–12: Model evaluation, optimization, and visualization of predictions Final Deliverables: Functional ML model with sample predictions Report and/or presentation highlighting insights and predictive trends Desired Skills and Time Commitment: Coding skills, organic chemistry, communication, and writing documentation of conducted work |
Phuong Cao (Computer Science) | Phuong Cao is a research scientist at the National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign. His research interests include securing high-performance scientific computing through large-scale measurements of network protocols, designing probabilistic models to detect cyberattacks, and synthesizing machine-assisted proofs of federated authentication protocols. His research has won a Best Paper Award, a Best Hackathon Award, and an Outstanding Mentor recognition for advising students in cybersecurity research. Phuong is also a Trusted CI Fellow of the NSF Cybersecurity Center of Excellence. Possible Project: AICyberLake: Building a Security Data Lake for Cyberattack Analysis Project Description The AICyberLake project curates a security data lake by sourcing cyberattacks from the DeltaAI system at NCSA and its peer supercomputing centers. The data includes Zeek network cryptographic metadata, GPU interconnect vulnerabilities, and ground-truth incident reports. The resulting data lake provides a real-time, anonymized stream of attack attempts to vetted research teams for evaluating their agentic AI-based detection models against unseen adversaries. Project Objectives The AICyberLake team will work with research groups to analyze attacks targeting U.S. supercomputing infrastructure and provide an API to share attack metadata with the broader community, including policymakers such as the National Institute of Standards and Technology (NIST). |
| Pranav Swaroop Cheluva Nagaraju (Civil and Environmental Engineering)  | Pranav Swaroop C N is a graduate student in Construction Engineering and Management at the University of Illinois Urbana-Champaign. He works with UIUC Facilities and Services, supporting construction management activities such as reviewing RFIs, shop drawings, and estimates to strengthen project coordination and design alignment. With a background in civil engineering, Pranav has teaching and research experience as a Civil Engineering Tutor and CAD Laboratory Teaching Assistant. He is also a national-level netball player and six-time state champion, experiences that honed his leadership and teamwork skills. His research on sustainable construction—using bamboo as reinforcement and bamboo leaf ash as a cement replacement—was recognized by the New Age Innovation Network (Government of India), earning ₹85,000 in funding. Pranav is eager to mentor undergraduates in the URSA program, supporting impactful projects and skill development in engineering research and innovation. Description of Possible Projects The two projects listed below may be adjusted based on interest or student needs but will consistently revolve around project engineering, construction management, and the use of Procore as a reference platform. Preferred Majors: - Civil Engineering - Construction Management - Architecture or Architectural Engineering Project 1: Construction Scheduling & Project Execution in a Project Management Framework Description: This project immerses students in the real-world workflow of a Project Engineer by guiding them through how construction schedules are developed and used within the project management tree—from owner and project manager down to project engineer and field supervision. Students will explore how phase schedules are created, tracked, and adjusted in response to field conditions and material or labor delays. Students will also gain insight into the role of scheduling in project execution, including how it connects with document control, trade coordination, and cost/time management. Procore will be used as a reference to show how schedules are maintained and tracked in professional settings. Key construction tasks integrated into the learning experience include the following: - Drawing and specification reviews - Daily site walks and documentation - Writing daily logs and phase updates - Managing change orders and communicating progress Preferred Majors: - Civil Engineering - Construction Management Weekly Commitment: One meeting per week for discussion, mentorship, and progress review. Timeline & Deliverables (May Change): Weeks 1–2: Introduction to project management hierarchy and project lifecycle Weeks 3–6: Build scope and phase schedule; simulate real-world changes Weeks 7–10: Analyze delays, simulate change impacts, align with field progress Weeks 11–13: Reflect on project coordination and reporting workflows Weeks 14–15: Submit summary report and final presentation Project 2: Construction Document Flow and Coordination Through the Lens of a Project Engineer Description: This project gives students a comprehensive look at the document handling responsibilities of a Project Engineer within a construction team structure. Students will walk through the entire flow of documents—from submittals and RFIs to closeout packages—understanding how these elements move across the project management hierarchy and contribute to project progress. Using Procore as a reference tool, students will simulate how Project Engineers: - Review and log submittals and shop drawings - Process and track RFIs and responses - Conduct onsite walks and create daily logs - Prepare and communicate weekly updates - Track change orders, document turnover, and project closeout - Understand how project completion is tracked using real-time systems This project reinforces construction communication, accountability, and document-based coordination from start to finish—mirroring professional expectations. Weekly Commitment: One mentorship session per week for process review, role-play, and feedback. Timeline & Deliverables (May Change): Weeks 1–2: Understand the PM team structure and communication pathways Weeks 3–6: Simulate handling of submittals and RFI workflows Weeks 7–10: Review field updates, logs, and track project status Weeks 11–13: Focus on change orders and closeout documentation Weeks 14–15: Deliver a project flow summary and role-based reflection Desired Skills: - Strong communication and organizational skills - Basic understanding of construction processes or eagerness to learn - Interest in project coordination, scheduling, and document management - Ability to commit to regular weekly meetings and complete assignments independently - Familiarity with general engineering or construction terminology is a plus but not required - Motivation to engage with project management workflows and tools like Procore - High interactivity and willingness to participate actively in discussions |
| Reva Baglane (Industrial and Enterprise Systems Engineering)  | Reva is a first-year Master’s student in Industrial Engineering at UIUC, driven by a passion for designing smarter, more efficient systems. With a Mechanical Engineering background and hands-on experience in automation, robotics, and resource optimization, she thrives at the intersection of engineering logic and real-world impact. Whether simulating biped robots or streamlining workflows, Reva approaches every challenge with curiosity, analytical rigor, and creative problem-solving. She’s especially drawn to systems that blend human-centered design with data-driven insights—be it in supply chains, operations, or emerging tech. Beyond academics, Reva is a community-minded leader, having served as a student council representative and led her university’s women’s volleyball team. A lifelong learner and natural collaborator, she’s committed to building solutions that work—for people, for systems, and for the future. Description of Possible Projects: Project Title: Leveraging IoT and Edge Analytics for Real-Time Inventory Monitoring in Warehousing Description of Possible Project: This project explores how Internet of Things (IoT) devices combined with lightweight edge analytics can be used to monitor inventory in real-time within a warehouse or small-scale storage environment. The system will simulate stock level monitoring using sensors (such as load cells, RFID, or ultrasonic sensors) connected to microcontrollers like Arduino or ESP32. Edge-based data processing will be used to reduce communication load and trigger alerts or decisions without needing continuous cloud connectivity. The ultimate aim is to create a responsive, cost-effective inventory tracking prototype that reflects how modern smart warehouses function. Target Skills and Majors: (Preferred but not required – an eagerness to learn and contribute is what matters!) Majors: Industrial Engineering, Mechanical Engineering, Electrical Engineering, Computer Science, Mechatronics Skills: Basic knowledge of Arduino or ESP32 programming Familiarity with IoT concepts Basic electronics and sensor integration Optional: Experience with data visualization or simple dashboards Timeline: Weeks 1–2 (Onboarding & Fundamentals): Introduction to IoT systems and edge computing. Student will explore microcontroller setup, sensor types, and basic communication protocols (e.g., MQTT, Serial, WiFi). Week 3 (Literature Review & Research): Study academic and industry applications of IoT in inventory monitoring. Review best practices in warehouse sensing and automation. Weeks 4–8 (Prototype Development & Testing): Design and assemble a physical prototype using sensors to detect item quantity or movement. Write embedded code to process data locally and trigger predefined actions (e.g., alerts when inventory is low). Week 9 (Experimentation & Evaluation): Test the prototype under different scenarios (e.g., varying inventory levels, sensor noise). Measure system accuracy, responsiveness, and data transmission efficiency. Week 10 (Refinement & Final Presentation): Fine-tune hardware/software as needed. Prepare a comprehensive final report and create a demonstration video or live demo. Submit code and documentation to GitHub. Desired Skills and Time Commitment: Skills: Arduino/ESP32, basic electronics, beginner-level coding (C/C++ or Python), analytical thinking Time Commitment: 8–10 hours per week (Includes hands-on prototyping, code testing, reading, and weekly check-ins) Outcome: Hands-on exposure to IoT systems in a real-world logistics context, technical experience in embedded systems, and an understanding of how smart supply chains use automation for decision-making. Desired Skills and Time Commitment: Skills: Arduino/ESP32, basic electronics, beginner-level coding (C/C++ or Python), analytical thinking Time Commitment: 8–10 hours per week (Includes hands-on prototyping, code testing, reading, and weekly check-ins) Outcome: Hands-on exposure to IoT systems in a real-world logistics context, technical experience in embedded systems, and an understanding of how smart supply chains use automation for decision-making. |
Savya Khosla (Computer Science) | Savya Khosla is a second-year Ph.D. student at the University of Illinois Urbana-Champaign, advised by Professors Derek Hoiem and Alexander Schwing. He earned his Master’s in Computer Science from UIUC and his Bachelor’s in Computer Engineering from Delhi Technological University. His research focuses on long-form video understanding and multimodal learning, with a focus on designing efficient visual representations. He has collaborated on research projects across academia and industry, working with researchers at leading institutions including Meta Reality Labs, Google Research, Adobe Research, the Allen Institute for AI, Mila, and the National University of Singapore. More information can be found at https://savya08.github.io/. Description of Possible Projects: Project Description: Have you ever tried finding a specific scene or object in hours of video footage—like finding Waldo in a crowd, but in a movie that’s 5 hours long? Current AI tools have made tremendous strides, but they fail to do this task efficiently. A big part of the reason is because they try to solve this by breaking every frame into a big grid of “patches” and searching through them all, which is slow and inefficient. This project aims to build a small but smart AI that focuses only on the important parts of each frame—like people, objects, or regions—so it can search faster and with less computing power. What You’ll Do: - Develop AI models to find and track meaningful regions in a video. - Build a simple search tool where a user can provide a query object, and the AI finds it in a long video or database. - Compare how fast and accurate your tool is compared to existing methods. Timeline & Deliverables: - Weeks 1–3: Learn the basics of video processing, set up the codebase, and test on small video clips. - Weeks 4–8: Implement the “region-based” search pipeline and run experiments. - Weeks 9–12: Compare results to a baseline method, and prepare a short presentation + demo video. Desired Skills and Time Commitment: Skills: - Required: Python programming - Nice to have: Experience with Pytorch Time Commitment: - 10 hours/week |