Hey, I'm Paul Guan

NLP/LLM - Copilosh / Copilosh WEB

December 2024

Copilo.sh is a wrapper function to add to your .bashrc or .zshrc file. Accompanied by a FastAPI local server, running a LM on CPU. It will catch all the errors you make in your terminal (non-zero exit code) and call the LM to generate a response to help you solve the error.

• Performance: Good but depends on the choice of the model
• Portability : Running with CPU and could be deployed in a Web Server
• Deployment : Github CI/CD + Microsoft Azur VM + Canary Deployment with Docker

Video of my project

Data Engineering - TermiCator

June 2024

This project, developed in Scala, aims to alert a large number of users efficiently and effectively. Key features and technologies used in this project include:

• Project Scalability : Designed to handle a high volume of alerts and user notifications.
• IoT Device Simulation : Simulated IoT devices triggering alerts based on specific criteria, ensuring relevant notifications are sent to the concerned users.
• Technologies : Utilized AWS Kinesis, Firehose, S3, DynamoDB, and EMR for data streaming, storage, and processing.
• Infrastructure as Code : Managed and provisioned resources using Terraform.

Some images of the project

Microsoft Azur - Weather

June 2024

This project focuses on predicting weather conditions using various Microsoft Azure services. Key components and technologies used in this project include:

• Custom Vision : Leveraged Azure Custom Vision for image recognition and analysis related to weather patterns.
• Machine Learning : Implemented machine learning models to predict weather conditions based on historical data and patterns.
• Azure Web App Service : Deployed the application using Azure Web App Service for scalability and ease of access.

Some images of the project

BigData - Displaying Stock Prices

April 2024 - May 2024

This project, developed in Python, aims to display stock prices for various companies. Key features and technologies used in this project include:

• Data Handling : Reading, cleaning, and storing large volumes of stock market data.
• Libraries Used : Utilized numpy and pandas for data manipulation and matplotlib for data visualization.
• Data Visualization : Displayed company stock prices using Dash for interactive and dynamic web applications.

Some images of the project

CNN - Image Classification

May 2024

This project, developed in Python using TensorFlow, aims to classify images of ships into 10 different categories. Key aspects of the project include:

• Goal : Building a convolutional neural network (CNN) to classify images of 10 types of ships.
• Technology : Developed the CNN using TensorFlow and Keras for efficient and accurate image classification.
• Model Training : Trained the models and evaluated their performance using relevant metrics.

Some images of the project

Multi-Agent - Simulating behavior in the metro

March 2024 - May 2024

This project aims to simulate people's behavior and study their reactions in the metro

• Technology : Developed in NetLOGO.

NLP - Restaurant Review Classification

April 2024 - May 2024

This project, developed in Python, aims to classify restaurant reviews. Key aspects of the project include:

• Dataset : Used the Semeval-ABSA dataset from Hugging Face.
• Data Processing : Tokenization using NLTK and regex, byte pair encoding, text normalization (stop words removal, lemmatization, conversion to lowercase).
• Data Statistics : Conducted comprehensive statistical analysis of the dataset.
• Model Training : Trained multiple predictive models including RNN and Naive Bayes.

Some images of the project

Hackathon - Ministry of Armed Forces

April 2024

Python project developped in 3 days that allows you to identify if a text is generated by an AI or human. Key aspects of the project include:

• Data : Training data with 2 labels (AI/Human)
• Algorithms used : Naives Bayesian/CNN/Logistic Regression

LinkedIn Posts

Sudoku Solver

March 2024 - April 2024

This project involves solving Sudoku puzzles using C#. Key aspects of the project include:

• Technology : Developed in C#.
• Algorithm : Implemented the Dancing Links method to efficiently solve Sudoku puzzles.

IDE Shrek

June 2023 - July 2024

This project involved developing an Integrated Development Environment (IDE) themed around Shrek. Key features and technologies used in this project include:

• Python Execution : Ability to run Python code.
• Frontend : Developed using JavaScript and React.
• Backend : Implemented in Java.
• Basic IDE Features : Create, open, delete, and move files.
• Advanced Features : Integrated Git and Maven functionalities.
• Deployment : Delivered as a Docker container for easy deployment and scalability.

Some images of the project

Tiger Compiler

March 2023 - May 2023

This group project, developed over the course of three months, aimed to create a basic compiler for the Tiger programming language using C++. Key aspects of the project include:

• Technology : Developed using C++.
• Group Collaboration : Successfully completed through effective teamwork and collaboration.
• Duration : Spanned three months of intensive development and testing.
• More informations about the project : https://assignments.lrde.epita.fr/

42SH

January 2023 - February 2023

This group project, developed over the course of four weeks, involved implementing a basic Unix shell command interpreter using the C programming language. Key aspects of the project include:

• Duration : Completed in four weeks.
• Technology : Developed using the C programming language.
• Functionality : Implementation of a basic Unix shell command interpreter.

Genetic Pipeline with Institut Curie

January 2022 - May 2022

This project, conducted in collaboration with the Institut Curie, focuses on recognizing cancer-causing variants using advanced computational techniques. Key aspects of the project include:

• Goal : Identification of cancer-causing genetic variants.
• Pipeline Development : Created a data processing pipeline using Snakemake.
• Neural Networks : Implemented neural networks using Keras for variant recognition.
• Bioinformatics Tools : Utilized bioinformatics software such as Varsim, Samtools, and IGV.
• Data Handling : Reading and writing data produced by bioinformatics software.
• Weekly Reports : Prepared and presented weekly progress reports.
• Deliverables : Authored comprehensive project deliverables.

Robotics Project with LEGO MINDSTORM

January 2021 - May 2021

This project involved designing and programming a LEGO MINDSTORM robot to collect balls, developed in Java. Key aspects of the project include:

• Programming Language : Implemented in Java.
• Team Collaboration : Executed as a group project.
• Weekly Reports : Prepared and presented weekly progress reports.
• Documentation : Authored various documents including project specifications, testing plans, and development plans.
• Project Defense : Delivered a final presentation at the end of the project.

Video of my project