ProjectsJustin Nguyen
Showing 19 of 19 projects
Low level functions from scratch
Implemented simple versions of low-level functions used in CNNs using Numpy, focusing on array manipulations and broadcasting.
Decision Tree Classifier from scratch
Implemented a decision tree classifier from scratch using Python and NumPy, featuring functions to grow the tree, calculate the Gini impurity, and grow trees using ensemble methods with bootstrap aggregating
Philosophising about Superintelligence
Wrote an essay on why consciousness is a prerequisite for true superintelligence, arguing that self-awareness and subjective experience are essential for advanced cognitive functions.
Automatic Sudoku Solver
A C# application that uses simple algorithms to solve Sudoku puzzles automatically and efficiently, providing a user-friendly interface for input and displaying solutions.
Camera geometric calibration & Simple image classification using CNN's
Calibrated a camera for 3D image reconstruction, additionally trained several CNN models for some simple image classification.
Exploring several Tree-based algorithms using a bank client dataset
Using tree-based models we classified clients based on whether they participated in an investment program or not.
Using Tree/Cluster-based algorithms to predict remaining treatment time for sepsis patients in hospital care processes
Applied machine learning algorithms to healthcare data for predicting treatment duration in sepsis patients.
Topic Modeling and Clustering algorithms to identify loneliness themes in Dutch COVID-19 tweets
Applied NLP techniques to analyze social media data and identify patterns of loneliness during the COVID-19 pandemic.
Automatic dog emotion recognition using Convolutional Neural Networks
Bachelor thesis project focused on training CNN models to recognize emotional states in dogs from image data.
Magnetoencephalography (MEG) signal classification of motor imagery movements
Advanced neural signal processing project using CNN to classify imagined motor movements from MEG brain signals.
Algorithms and Data Structures from scratch
Coded fundamental algorithms and data structures such as binary search, quick sort, and linked lists in C#, focusing on efficiency and performance.
Time series forecasting on temporal laser sensor data using LSTMs and GRUs
Implemented LSTM and GRU models to forecast time series data, focusing on sequence prediction and temporal patterns.
Creating a keyboard from scratch
Designing and building a custom keyboard PCB using mechanical switches and microcontrollers to create a personalized typing experience.
Classifying spam using simple machine learning classifiers
Implemented various machine learning algorithms to classify spam messages using text processing and feature extraction techniques.
Retrieval Augmented Generation (RAG) chatbot for 510 - Red Cross
Implemented a RAG chatbot using LangChain and OpenAI to assist with answering questions for Ukrainian refugees, enhancing response accuracy with document retrieval.
Simple dashboarding web app using Flask and Plotly Dash
Created a simple web application using Flask and Plotly Dash to visualize data interactively with charts and graphs.
Detecting behavioral patterns with sensor data.
Using the data of Time of Flight (TOF), Thermopile (TMP) and Inertial Measurement Unit (IMU) sensors to detect patterns in human behavior.
Text-based dialog system
Creating a traditional state-based dialog system using Python for natural language interaction and conversation management.
Finding the Most Valuable Parts of a Dataset
Developed and evaluated three Spark-based methods for identifying and retaining the most valuable subset of relational datasets using distributed computing.