Projects
Key projects and key points. Coursework and personal projects.
Personal Project
Full Stack Developer - AWS/Vercel Hosted Resume Website
Mar '24 - Current
A resume-based site using a Next.js front-end server on Vercel and a Spring back-end server on AWS.
- Created a dynamically updatable front-end server with proper resource caching and hosted it on Vercel
- Designed a database for the front-end and implemented it using PostgreSQL along with an interface in Java using Spring Framework
- Hosted the back-end on AWS RDS + AWS EC2 and created a static resource bucket on S3 with caching on the edge
- Implemented an Auth.js-based middleware layer to secure front-end and complemented it with a Spring JWT role-based authentication scheme on the back-end
- Mapped proper DNS entries on Namespace and created proper SSL certification to prevent data leakage and CORS errors between the front-end and back-end.
Capstone
Full-Stack Developer - Next.js + Spring Storefront
Dec '23 - Mar '24
An example storefront using a Spring-based back-end and Next.js front-end.
- Implemented a Spring-based REST API to handle and serve user data, user preferences, and product + category data.
- Modeled the database using UML ensuring desired product functionality was met and used MySQL to realize the database.
- Created custom SQL queries within the repository to handle operations which could not be inferred by the framework.
- Designed a UI framework by hand using Photoshop and created a Typescript implementation using a Next.js based server with Tailwind classes.
Coursework
ML Engineer - Object-Based Occlusion Inpainting
Sep '18 - Dec '18
An attempt to create a pipeline for organic occlusion inpainting.
- Created a deep learning pipeline designed to identify organic occlusions and inpaint the localized object in a team of four.
- Created a dataset with organic occlusions and transfer trained a YOLO network to detect the objects and return bounding boxes for each occlusion.
- Created a lambda layer to convert bounding box outputs into a rough segmentation map for use in the inpainting network.
- Learned how to use Docker to resolve issues that arose following Python pipeline Dockerization issues.
Coursework
GPU Software Developer - GPU Optimized Face Detection
Sep '18 - Oct '18
Cuda C++ implementation of the Viola-Jones Detection algorithm
- Designed and implemented a GPU optimized version of the Viola Jones detection algorithm in a team of three.
- Implemented the variable size image pyramid generator and the Adaboost algorithm used for training using CUDA C++.
- Presented on how to parallalize the Adaboost algorithm by taking into account GPU architecture constants.
- Reported and fixed GPU memory management issues using the Visual Studio GPU debugging tools.
Internship
Researcher - Endoscopy Robot
May '18 - May '19
Summer Research Internship Project
- Researched and designed a control scheme for a colonoscopy robot and implemented the communication packet structure using ARM C to triple initial segmented robot polling rate
- Implemented a variable segment-based GUI using Python and MATLAB to collect fused data and individually identify robot pose of up to five connected robot segments.
- Collected robot movement fusion and orientation data, compiled it within a scientific journal paper, and prepared a presentation on the prototype and research progress for an engineering outreach event.