Hey, I'm Siven!
I'm an computer engineering and computer science student at Duke University. I write software and create systems to help people, whether it be in health, education, or software development. I also take pictures sometimes.
Skills
I program in a number of languages and with a broad range of principles, from frontend and design to backend, to simple scripts. Most of my work is done in Java and TypeScript, but I have worked extensively with Go, Python, C, and C++
Favorite/Frequent Languages and Tools
Languages
Technologies
Web Development Frameworks
Miscellaneous
Experience
Duke University Dept. of Biostatics & Bioinformatics
Working with multi-omic and genetic sequence data to find associations in the genome. More to come!
Oklahoma Medical Research Foundation
Experimented with Transformer and Convolutional Machine Learning models in Python to determine the viability of a feature reduction model over an entire genomic track. Included building custom tokenizers to convert research standard BED files into vectors.
Red River Community Corps
Designed and architected software for curriculum delivery systems alongside Slalom Consulting, as well as contributed to an AWS Imagine Grant. Designed workflows for student podcast and media creation, and custom software tools to automate these processes utilizing tools such as ClickUp. Additionally built volunteer management tools that linked external databases and volunteer management solutions.
Projects
LARI
EHR-integrated application that automates managing patient appointment scheduling (and rescheduling).
VeloceQuiz
Utilized Large Language Models and text classifiers (OpenAI API and LangChain) to create a web resource for students to generate practice tests from a PDF of notes. Text classification via LangChain allowed for users to select a number of questions from particular generalized concepts generated from the notes.
Soundbits
Architected, designed and built a social media application that allows users to share and send "bits" of sound that they come across in their daily life, and enable other uses to use them as a library of samples for them to use in their music productions.
FIRST Robotics Match Optimizer
Built matchup optimizer utilizing Random Forest Classifiers and data pulled from official FIRST APIs to train a Random Forest Classifier alongside Binary Classification to predict wins to about an 85% accuracy rate. Included a Svelte frontend and visualization to inform match strategy.
roundify
Customized user interface for Firefox (via userchrome.css), includes integration that modifies the default homepage with the daily Astronomy Picture of the Day.
mtxmathtools
CLI matrix computational tool I wrote to make doing linear algebra homework easier. Wrapped in a python package.