home
automation projects
pass genius

pass genius
chrome-extension

A minimalist chrome extension that helps users generate strong passwords on the fly. It takes into account preferences such as lower case letters, upper case letters, digits, and special characters. The extension is now available on the chrome store.
git spawn

git spawn
shell-script

Git spawn is a macOS automation project utilizing Python scripts. It empowers users to effortlessly create GitHub repositories on-the-fly directly from their Mac's Spotlight. With streamlined efficiency, Git Spawn enhances the repository creation process, providing a seamless and time-saving experience for developers managing numerous projects.
resume sync

resume sync
shell-script

Resume sync is an automation tool that keeps your developer portfolio seamlessly updated with your latest resume. Leveraging shell scripts, it monitors changes to your resume file and automatically deploys the updated version to your online portfolio. No more manual uploads - ResumeSync ensures your credentials stay current. Its intelligent file monitoring detects modifications and triggers a streamlined deployment process, syncing the new resume version without intervention.
other projects
urban mobility analysis dashboard

urban mobility analysis dashboard
d3.js

The project involved creating an interactive dashboard with 7 visualizations using D3.js and SQLite schemas to analyze urban mobility and lifestyle trends in Ohio. Project management for a six-member team followed Agile methodology, emphasizing strong communication skills. User experience was improved with tooltips, legends, and opacity adjustments for enhanced chart details and selection clarity.
beeswarm visualization of global development trends

beeswarm visualization of global development trends
d3.js

The project involved creating an interactive beeswarm chart using D3.js to visualize global development data, enabling users to explore trends across countries and regions over time. Smooth animated transitions and staggered delays were implemented to intuitively convey data changes. A responsive user interface with filtering controls for attributes like GDP and population dynamically updated the visualization in response to user input, enhancing usability.
scalable image classification

scalable image classification
aws

The project focused on enhancing system performance and scalability by optimizing auto-scaling policies and CloudWatch alarms, and implementing custom step-scaling based on m1-m2 metrics, resulting in a processing time reduction of over 50%. Additionally, a scalable AWS architecture was implemented for image classification using Python/Boto S3 and Java/AWS SDK, enabling efficient processing of over 100 image requests with an average processing time of less than one second.
hybrid cloud using openstack and aws

hybrid cloud using openstack and aws
openstack

I developed a hybrid cloud application utilizing OpenStack and AWS services like EC2, S3, Lambda, and DynamoDB for video processing. Reduced infrastructure costs by 30% with an OpenStack private cloud. Created a docker container with face recognition libraries for Lambda function, cutting cold start time by 40%. Implemented OpenStack private cloud with core services using Dev-Stack on Ubuntu VM.
cluster validation - meal data analysis

cluster validation - meal data analysis
python

The project involved building an end-to-end data pipeline in Python for clustering and analyzing over 15,000 meal glucose readings to uncover trends in daily carb intake based on CGM and Insulin data. Investigation into DBSCAN and KMeans algorithms revealed DBSCAN had 20% higher accuracy for irregular meals. An automated pipeline was implemented for large-scale analysis to uncover personalized carb intake patterns from glucose data, including preprocessing, feature engineering, and clustering with KMeans and DBSCAN to group meals into seven carb categories.
lambda-powered face recognition

lambda-powered face recognition
aws

The project involved architecting a serverless application on AWS using Lambda, S3, and DynamoDB for real-time face recognition, reducing costs by 70% compared to an EC2-based approach. One-click app deployment and rollback were enabled using Docker and GitHub actions. Additionally, a load testing framework was crafted to simulate real-world traffic, guaranteeing processing of 100 sample videos in under 10 minutes, even during peak loads. Containerization was leveraged to slash cold start times by 40%.
handwritten digit classification

handwritten digit classification
android

I developed a mobile application for handwritten digit recognition, utilizing one master smartphone and four slave server devices for distributed processing. Optimized image segmentation and distributed ML models across devices resulted in a 50% reduction in image processing time and 75% faster predictions respectively, while improving inter-device communication by 60%.
fibonacci calculator

fibonacci calculator
aws

The project involved launching a web application on AWS Elastic Beanstalk (EBS) that calculated the Fibonacci series up to a number and cached the result in a Redis database, significantly improving search speeds by 98%. Results were stored in a PostgreSQL database to maintain a record of subsequent searches.