Presentation Schedule

HEARTNet is a deep learning–based framework designed for real-time MI detection and high-resolution cardiac activation reconstruction from surface ECG signals. 

This project utilizes ML to optimize cleaning time in membrane bioreactors by analyzing transmembrane pressure trends. The system reduces downtime, energy use, and unnecessary cleaning, which enables more efficient membrane use and control.

We use novel methods to explain and change the outputs of ProteinMPNN, a popular protein design model. We use it to stabilize antibodies used to treat disease. Stable antibodies have a longer half-life in the body and are effective at lower doses.

This project leverages hydrogels to develop a sustainable, ethical, and cost-effective alternative for the medical device industry, reducing dependence on animal blood in deep vein thrombosis research and testing.

AI-guided system that optimizes stepwise differentiation protocols to guide iPSCs into neuronal organoids that enhance precision, diversity, reproducibility, and can replace the current method of preclinical trials.

Our project develops a multilayer biosensor for continuous hydrogen peroxide monitoring in interstitial fluid as a marker of oxidative stress. Our work focuses on electrode optimization, benchtop electrochemical validation, and integration with microfluidic channels, microneedles, and prototype electronics for minimally invasive sensing.

Reverse directed evolution strategy to engineer a smaller, more specific Cre by systematically removing non-essential regions while preserving the catalytic domains required for DNA cleavage.

Our project focuses on genetically engineering exosomes for glioblastoma-cell targeting. We are designing them to become a multi-gene system that targets three GBM biomarkers, making them an efficient and selective vehicle for cancer drug delivery.

PeriDevice is a custom SLA-printed packed-bed bioreactor that enhances peritoneal dialysis by regenerating dialysate through alginate-based microcapsules containing urease to shorten the number of cycles, addressing key engineering challenges, including the tradeoff between flow rate and pressure drop within the system and the stability and mechanical integrity of the capsules.

This project develops an at-home, point-of-care device for postpartum depression screening. Using saliva-based hormone detection, paper microfluidics, and smartphone analysis, the system enables simple, low-cost, postpartum-focused health monitoring.

Water reuse is a crucial step in conserving water in California. This team redesigned an advanced water purification system for the City of Palo Alto to treat effluent from the city’s wastewater treatment plant and promote recycled water reuse.

Communities in rural Mexico need improved housing structures that use bamboo, an accessible and plentiful natural resource. This project designs structural connections using both 3D printing and PVC materials.

The community of Pajaro has been dealing with a serious risk of flooding, due to an undersized levee system along the Pajaro River. Our goal is to redesign their flood protection system for a 100-year storm, taking into account climate change.

Can you imagine SCDI turned into wood? But not just any wood - huge engineered wood products, called Mass Timber - look it up! This project will redesign SCDI into Mass Timber and analyze the impact to sustainability, cost, and time.

Based on a proposed self-sufficient logistics center in Bakersfield, this project addresses design and construction strategies to create an efficient, sustainable site while maximizing benefits for local communities and site users.

Silica nanoparticles of various sizes and concentrations were fabricated and added to self-consolidating concrete to determine their effects on concrete strength and environmental impact.

We will be competing in the 2026 ASCE Concrete Canoe Regional Competition (MidPac) taking place April 10-12 2026.  We are creating a unique concrete mix and hull design in order to maximize both speed and buoyancy of this canoe.

Redesigning Malley to have a larger area will help bring in more students. Modernizing the look and utilities can be used to make SCU’s gym a selling point of the school instead of just a part of it.

This team designed and fabricated a steel bridge emphasizing structural efficiency and constructability. The team collaboratively handled analysis, scheduling, detailing, and erection planning while prioritizing competition rule compliance.

Using existing machine learning models within the speech diarization realm, we plan on improving diarization between parent and child conversations. Current models work well with adult-to-adult speech but struggle for child-directed speech.

A high-performance, web-based version of Street Fighter III that evaluates different LLM architectures through gameplay against trained AI agents.

JSD is a defense apparatus against privacy attacks using synthetic data, combining a state-of-the-art implicit-privacy detection model with a genetic algorithm dataset optimization pipeline.

EcoVisor is a web-based AI tool that transforms complex sustainability articles into clear, data-driven visuals. It bridges the gap between technical analysts and the general public, making environmental insights accessible and easy to understand.

A note-taking software that helps create and edit notes in a network (resembling a neural network) that ties ideas and notes together for better information indexing, retention, and retrieval.

Student portal for finding campus events, integrating with Google Calendar, and tailoring events by interest, while supporting a user-friendly moderator backend incorporating human-AI collaboration. 

BiKi is an interactive, multilingual math learning app designed for bilingual kindergarteners. It uses visual and auditory aids in activities to bridge math vocabulary between English and students’ native languages.

A wellness app focusing on sleep health for students and professionals through data-driven insights. Using wearable technologies, our app delivers smart alarms, sleep analysis, and AI-powered recommendations to support healthier daily productivity.

An AI-powered wellness platform that helps users build personalized fitness and nutrition plans using community-shared routines and recipes. Designed for individuals at any starting point, including beginners and those with physical limitations, the platform combines real user experiences with personalized AI-generated options to make healthy living accessible and sustainable.

Bouillean is a text-to-text recipe generation model that mimics how chefs approach recipes by reducing recipes into standardized components and constructing dishes based on those components.

Fast and intelligent entity linking and resolution for real, messy datasets.

A web app for SCU student organizations and the Center for Student Involvement that centralizes resources such as club registration and financial management to improve the efficiency and visibility of student organizations for SCU students and staff.

Kaikei centralizes club finance by unifying interactions and automating budget tracking. It streamlines treasury duties, ensuring fiscal clarity and reducing administrative overhead for student organizations.

Redesigning the IAJES website into an interactive hub for discovering members and collaborating across Jesuit engineering schools worldwide. Built with React and Supabase, it’s accessible, maintainable, and scalable.

A low-cost, autonomous delivery rover designed to deliver food and drinks to workspaces on campus. The system uses ROS2, NAV2, and LiDAR-based perception to safely navigate a fixed route.

Our project implements a triangle sorting-based, real-time 3D rendering pipeline. We utilized this renderer to implement a game engine, featuring camera control, user input, audio feedback, etc.

An interactive game that transforms sustainability education into a collective community experience. It replaces passive learning with a simulation where users practice real-world sustainable decision-making.

Our project strives to enable musicians to protect their work from being stolen for use in GenAI training data.

BridgeIDE is a cloud-based collaborative IDE built specifically for CS students and educators. It provides pre-configured, course-ready environments run entirely in the browser, allowing users to code, compile, and execute assignments instantly.

The Interactive Debate Simulator is an AI-powered platform providing real-time voice debates and objective feedback to help learners master argumentation and public speaking at scale.

We present an implementation of a mixed-precision iterative refinement solver leveraging collaborative execution between the CPU and a Neural Processing Unit (NPU) within an AMD Accelerated Processing Unit (APU).

Declining literacy rates, sociopolitical turmoil, and rapidly evolving technologies threaten to uproot the foundations of human subjectivity and creativity as we know them. The need for such a design space to help people exercise more intellectual rigor towards general concepts and values has never been greater.

Creating a probe for early detection and neutralization of the invasive species Limnoperna Fortunei, aka the Golden Mussel. Also creating a dashboard for government officials that will live ping locations of Golden Mussel detections.

We developed a high-resolution machine learning precipitation model for Santa Clara. By downscaling global forecasts (HRRR) using local terrain data, we aimed to deliver neighborhood-level accuracy superior to standard baselines.

Low-resource mobile and web platform combining lightweight predictive analytics, GIS mapping, and LiDAR terrain models with community data. Designed for low connectivity and to integrate traditional farming knowledge for Yucatán farmers.

Fisheye cameras are underutilized in traffic surveillance due to the scarcity of public annotated datasets. We reimplement the state-of-the-art model for fisheye traffic perception and explore improvements to detection accuracy and robustness.

Corridor counting is the process of determining the number of vehicles that traverse through a predetermined corridor. We create a real-time compatible system to accurately count vehicles utilizing deep learning methods on edge devices.

Aimed at assisting the visually impaired, our device pairs wide-angle camera vision with AI-based object detection to identify a range of potential obstacles, including rapidly approaching or elevated hazards. When a hazard is detected, haptic feedback alerts the wearer to its position.

A compact, pocket-sized language translation device that features a variety of different language combinations, voice cloning, and fluid conversation support.

This project builds a low-power smart agriculture monitoring system using 6TiSCH networking. Wireless sensor nodes collect soil and environmental data and transmit it over a time-synchronized IPv6 mesh network to a cloud database with a dashboard.

In this project, we design and validate a prototype reconfigurable RF filter, using microstrip geometric designs with electronically controlled tuning mechanisms. This component will allow for the real-time adjustment of filtering characteristics, enabling more dynamic wireless communication.

To design, fabricate, and validate a Current Mode DAC with the help of Texas Instruments and Cadence Design Systems. To comprehensively document all steps taken throughout the project for a future tape-out course at Santa Clara University.

Our project recreates & enhances the MIT Coffee-Can radar to build a modern, modular sensing platform. The system produces a chirped signal, transmits it, & processes echoes to measure range & motion, supporting hands-on RF learning & experimentation

This project classifies radio frequency devices by training a CNN to detect hardware-induced imperfections in RF signals. These imperfections form device-specific RF fingerprints that enable reliable device identification and classification.

SCRAP V enhances SCU’s radio telescope system by developing a robust MATLAB-based data processing and visualization infrastructure. Through this, it will enable accessible, real-time radio astronomy learning and research for students.

This project involves developing a handheld real-time positioning system that combines IMU and RFID sensors with machine learning to accurately track device position in 3D space. The system utilizes low-cost sensors to provide spatial awareness in different indoor settings.

This project builds a ROS2-based AI smart robot car that integrates vision, speech interaction, and autonomous control, providing an educational platform to demonstrate the complete perception–decision–execution pipeline on real hardware.

This project builds a real-time basketball video detection system on an FPGA that accelerates a lightweight vision model to detect players and the ball, achieving lower latency and higher throughput than a software-only implementation.

A spherical robot with a magnetic head that rolls, climbs, and explores tight, hazardous spaces humans can’t reach—ideal for inspection, construction, and search-and-rescue.

This project aims to work on an open-source implementation of the ARM Cortex-M0 processor in Verilog for research and experimentation.

This project involves the development of aptamer-based biosensors for the detection of E. coli and S. enterica via biochemical immobilization of aptamers on biochips for food safety & control purposes.

Transcranial photobiomodulation (tPBM) is an experimental treatment for neurodegenerative disorders and neurological conditions. NeuroGen is a photobiomodulation device designed to optimize light-stimulation therapy based on electroencephalography (EEG) readings. We will use it to test the effects of tPBM on cognitive performance and the associated EEG biomarkers.

An autonomous platform capable of conducting cellular biology experiments in a well-plate within a CubeSat environment. Biopayload will initiate experiments, monitor them, and collect data without human intervention in microgravity.

We are utilizing AI to improve the differentiation efficiency of iPSCs into retinal pigment epithelium cells (RPE), which are critical to vision. The AI model is trained to predict the optimal amount of growth factors & timing to increase accuracy. 

IRIS is an autonomous marine telemetry system designed for near-real-time data transmission from subsea sensors. Using a winch, the system periodically deploys a surface buoy to establish a communication link. This method replaces manual diver-based retrieval, allowing for persistent monitoring of benthic environments with minimal human intervention.

This project partners with Oorja to design a low‑cost, multi‑outlet water delivery system for solar‑powered irrigation. The system integrates modular piping, flow measurement, and simple controls to enable efficient and equitable water distribution, reduce labor demands, and improve productivity for smallholder farmers in rural Indian communities.

Design and implementation of a thermal imaging reconstruction system that generates three-dimensional building models from infrared imagery for energy efficiency diagnostics and building performance evaluation.

A low-cost multi-channel, portable sampling system for filtering environmental DNA from bodies of water. Analyzed samples can be used in impact assessments, biodiversity studies, and tracking invasive or endangered species. 

The problem addressed is how SCU can implement a user-friendly smart waste management system that improves waste sorting accuracy, reduces contamination, and provides real-time data to support more efficient waste collection practices.

To develop a low-cost, user-friendly Autonomous Underwater Vehicle (A.U.V) based ocean health monitoring system capable of collecting water column data at predetermined waypoints across a body of water.

HALoN strives to provide an accessible, crisis-aware emergency communication system to provide connection in times of disaster. It utilizes a LoRa mesh network, reconfigurable antennas, and a light-weight machine learning scheduling algorithm.

The Gesture-Controlled Reminder Device is a small, wearable device that manages reminders without screens. Hand gestures trigger haptic, light, and audio feedback, enabling hands-free, accessible task management for users who struggle with reminders.

Our project aims to reduce occupational strain with an adaptive, lightweight neck exoskeleton. It preserves full mobility while using real-time sensing and feedback to reduce fatigue and prevent long-term neck injuries in dynamic work environments.

A walker with instruments that measure gait (characteristic of walking) for clinicians. This data will help clinicians monitor and modify their patients treatment plan over a long stretch of time.

Our project is a dynamic resistance cable system that is designed to make physical therapy for the shoulder smarter and more responsive to real-time changes. It will use a motor and sensors which constantly adjust based on how much force is applied.

This project integrates computer vision, data collection, and robotics to build an intelligent PCB defect inspection system, enabling real-time defect classification and autonomous PCB retrieval and sorting using a UR3e robot.

A multimodal ergonomic risk assessment system that integrates computer vision, wearable electromyography (EMG) and inertial measurement unit (IMU) sensors, and a collaborative robot (COBOT) to monitor and reduce musculoskeletal disorder risk in real-time occupational settings.

Design of a robot dog for the purpose of disability aid and assistance. This is the first year of a multi-year project, which focued on creating the design parameters based on consumer needs and the mechanical design and basic control of the legs.

Our project aims to develop a modular, snake-inspired robot capable of versatile movement. The system will use locomotion algorithms and camera-based sensing to enable autonomous navigation.

The project is an extension within Blender that aims to clean markerless motion-captured data. It focuses on eliminating negligible key frames, cleaning jittery frames, locking down feet, and rescaling motion onto various proportioned models.

StudyScape is an app designed to help Santa Clara University students make decisions about where they study across campus. Providing real-time information such as noise level, occupancy, and overall activity.

In this project, we implement image processing on a PYNQ‑Z2–based autonomous system using a convolutional neural network, comparing a hardware‑accelerated FPGA approach with a purely software‑based solution.

A ROS2-based distributed control system that enables coordinated multi-robot aggregation and formation control in shared environments. The project implements and translates a simulation-derived control algorithm into a real-time, deployable multi-agent system.

This machine will be a table tennis serving machine that costs around $250 and with a small and lightweight design for a backpack. Functions will include: shot location, speed, and spin. The balls will be launched by contact from three flywheels.

A.T.L.A.S. is a roof-mounted single-axis solar tracking platform with an automated sprinkler-based self-cleaning system. By following the sun and keeping panels debris-free, it targets 20-50% higher energy output than fixed residential installations.

Our senior design project focuses on developing a grip-strengthening rehabilitation device for post-stroke patients with limited hand mobility and reduced grip function, allowing them to do their therapy from home. 

This project entails building a wirelessly-charged battery-powered light rail locomotive. The inductive charger will deliver a high power density to the train, enabling station stop times of one minute for a full-scale light rail system.

The SCU DBF Aerodynamics Team will design, build, and fly a model aircraft for the AIAA Design/Build/Fly competition, using aerodynamic theory, CFD, and experimental validation to optimize, lift, drag, and stability.

Design, analyze, and manufacture all structural components of an RC aircraft, optimizing strength, weight, and payload capacity to meet mission requirements and maximize performance for international competition. 

The objective of this project is to continue development of a compact and cost-effective vacuum heat furnace capable of achieving elevated temperatures for laboratory testing of material properties.

We are designing an autonomous eVTOL drone for wildfire response that provides rapid aerial situational awareness and payload delivery, improving early decision-making when ground access is limited.

Designed a pool filtration lab test rig featuring adaptive pump control. The system leverages real-time sensor data to enable demand-response energy savings and predictive maintenance, optimizing both cost and grid support.

Our project goal is to replace traditional camshaft valve actuation with variable pneumatic valve actuation to improve internal combustion engine power and efficiency.