Projects

Proposed a flexible and lightweight scheduler for network bandwidth allocation to approximate the ideal Weighted Fair Queuing(WFQ)
Implemented the hierarchical scheduler with C++
Applied NS2 and FCT topology to conduct the simulation
Carried out the evaluation and visualized data with the tools of Pandas, NumPy, Matplotlib

Implemented a Machine Learning model for predicting spam SMS message on AWS cloud service
Built and trained a spam filter machine learning model using XGBoost with Amazon SageMaker instance and deployed the model to an endpoint
Created Lambda function to trigger the filter when any new email file came
Formulated an AWS CloudFormation template for the system to represent all the infrastructures
Github: (https://github.com/Kathy961225/CloudComputingSpring2020-HW4)

Retrieved datasets from several sources to dig some useful information relating to COVID-19 with big data theory
Investigated on research paper analysis with Pyspark tool to figure out which kinds of people are at the highest risk getting infected
Implemented pandemic trend prediction with Pyspark ML library to forecast with the historical data
Establised the relationship between COVID-19 and age, gender, weather, certain regions, social health condition, community mobility, using Pyspark dataframe and Pyspark SQL

Implemented a photo album web application which could be searched using natural language through both text and voice
Used AWS S3 bucket to host the frontend and store the photos
Applied Amazon Transcribe to transcribe speech to text for searching
Created and trained Amazon Lex bot to handle the input queries
Deployed the project via AWS CodePipeline and launched service within inside VPC to prevent unauthorized internet access
Github: (https://github.com/Kathy961225/CloudComputingSpring2020-HW3)

Implemented a smart door authentication system, through which owner could give the access to visitors
Use the Kinesis Video Stream and Kinesis Data Stream to process the vedio gathering by the camera
Applied Rekognition tool to detect face
Implemented Lambda function to process the queries and hook different compoents and used API Gateway to manage the API
Github: (https://github.com/Kathy961225/CloudComputingSpring2020-HW2)

Implemented a serverless, microservice-driven web application with Natural Language Processing-powered application to book and search the restaurants’ information in Manhattan, NYC
Hosted the frontend in AWS S3 bucket and used API Gateway to manage the API
Fetched the restaurants’ data by Yelp API, stored the data in DynamoDB and indexed by AWS ElasticSearch Service to improve the search efficiency
Built the chatbot with Amazon Lex to generate intents and gather information from user side
Created Lambda functions triggered by CloudWatch to handle queries through SQS and sent reservastion results to user through SNS service.
Github: (https://github.com/Kathy961225/CloudComputingSpring2020)
Vedio: (https://youtu.be/lEmKRVWnb6U)

Implemented a complete image content positioning and text recognition prototype system, which was specifically applied to the license plate automatic identification system
Involved in 4 main modules including license plate detecting (OpenCV), license plate positioning (binarized image), character segmentation (image sliding windows and CNN) and Character recognition (Inception Net V3 module)
Realized auto location and recognition of license plate through the combination of deep learning framework TensorFlow, the machine learning algorithm and the image recognition algorithm
Achieved recognition accuracy of 80.0128%