ESP-32-controlled, wireless networking, IoT device that measures the occupancy of indoor rooms by probing WiFi traffic. Deployed as a web app (using DynamoDB, AWS IoT Core, Python Flask/Dash, MQTT Protocol, C++).
Won runner-up for best ECE Senior Design project.
A CUDA/GPU-optimized Convolutional Neural Network (CNN) for labeling images in the fashion MNIST dataset, improving to a ~86-88% classification accuracy and ~2.2x faster runtime. Utilizes Nsight/Nsys for performance profiling
OpenCV-based project leveraging Mask-RCNNs, Shi-Tomasi corner detection, SAM (Segment Anything Model)/instance segmentation, YOLO, and camera calibration to train, identify, and localize chess pieces and board cells. Uses Stockfish for chess game analysis
Linux-based kernel + operating system (built from scratch). Supports paging, multiple system calls, exception handling, interrupts, task running, multiple terminals, scheduling, keyboard, RTC, file system (written in C and x86)
CRUD/SQL-based web application and dashboard that allows users to query movies from the IMDB database and add, edit, and delete reviews. Written in Angular (TypeScript), Node.js, and hosted on Google Cloud Platform
C++, command line-based program that determines the shortest flight route between any airport in the world utilizing graph/path-finding algorithms (Dijkstra’s algorithm, A-star search, and breadth-first traversal) and the OpenFlights dataset
Web application that provides automatically updated time-series visualizations and future statistical projections on COVID-19 in the United States, utilizing Python’s Plotly, Bokeh, Dash, Pandas, and Matplotlib libraries, and deployed on Heroku cloud