Coursework Projects

Face Recognition with FaceNet

Built a face recognition system using FaceNet embeddings to perform identity verification and recognition.

Course: Deep Learning Specialization – Course 4, Week 4

🧠 Identity Verification Using Deep Face Embeddings

Image Segmentation with U-Net

Implemented a U-Net architecture for road segmentation from dashcam images using pixel-wise classification.

Course: Deep Learning Specialization – Course 4, Week 3

🧠 Semantic Segmentation for Road Scenes with U-Net

Car Detection with YOLO

Used a pre-trained YOLO model to detect cars in images and video frames using non-max suppression and bounding box visualization.

Course: Deep Learning Specialization – Course 4, Week 3

🚗 Object Detection using YOLOv2 & Non-Max Suppression

Transfer Learning with MobileNet

Applied a pretrained MobileNet model for image classification, using feature extraction and fine-tuning for edge-efficient inference.

Course: Deep Learning Specialization – Course 4, Week 2

📱 Lightweight CNNs for Mobile & Embedded AI

Residual Networks (ResNets)

Implemented a deep residual network (ResNet) with skip connections to solve vanishing gradient issues and build deep CNNs.

Course: Deep Learning Specialization – Course 4, Week 2

🔁 Skip Connections for Training Deep CNNs

Convolutional Model Application

Applied a pretrained CNN using transfer learning to classify images into categories with Keras and TensorFlow.

Course: Deep Learning Specialization – Course 4, Week 1

🖼️ Real-World Image Classification using Transfer Learning

Convolutional Model Step-by-Step

Manually implemented CNN architecture from scratch including forward and backward passes for convolution, pooling, and dense layers using NumPy.

Course: Deep Learning Specialization – Course 4, Week 1

🧠 Low-Level CNN Components Coded from Scratch

TensorFlow Programming Framework

Introduced core TensorFlow concepts like tensors, sessions, placeholders, and computation graphs for simple model training.

Course: Deep Learning Specialization – Course 2, Week 3

🔧 Build Simple Models using TensorFlow Basics

Optimization Methods

Implemented and compared mini-batch gradient descent, momentum, and Adam optimizers for efficient deep network training and convergence.

Course: Deep Learning Specialization – Course 2, Week 2

⚡ Optimize Deep Learning with Momentum & Adam

Gradient Checking

Implemented numerical gradient checking to validate backpropagation in deep neural networks by comparing analytical and approximated gradients.

Course: Deep Learning Specialization – Course 2, Week 3

🧪 Validate Backprop with Analytical vs Numerical Gradients

Regularization Techniques

Implemented L2 regularization and dropout to prevent overfitting in deep neural networks, improving generalization and reducing test error.

Course: Deep Learning Specialization – Course 2, Week 2

📉 Prevent Overfitting with L2 & Dropout Regularization

Initialization Techniques

Explored Zero, Random, and He initialization methods and their impact on gradient behavior, training speed, and convergence in deep neural networks.

Course: Deep Learning Specialization – Course 2, Week 1

⚙️ Comparing Weight Initialization Methods in DNNs

Deep Neural Network Application

Implemented a deep L-layer neural network to classify images as cat or non-cat, applying full forward and backward propagation, cost optimization, and prediction using NumPy.

Course: Deep Learning Specialization – Course 1, Week 4

🐱 L-Layer DNN for Binary Image Classification

Building a Deep Neural Network

Constructed a modular L-layer deep neural network from scratch using only NumPy, including all core functions from initialization to backpropagation and parameter updates.

Course: Deep Learning Specialization – Course 1, Week 4

🧠 Layered Architecture Built from Scratch with NumPy

Planar Data Classification

Built a shallow neural network with one hidden layer to classify non-linearly separable 2D data using NumPy, including full forward/backward propagation.

Course: Deep Learning Specialization – Course 1, Week 3

🔮 Shallow Neural Network for 2D Nonlinear Classification

Logistic Regression with Neural Network Mindset

Built a binary image classifier using logistic regression and NumPy with a neural network perspective. Applied vectorized computation and gradient descent to train the model on cat/non-cat data.

Course: Deep Learning Specialization – Course 1, Week 2

📸 Binary Image Classifier with Vectorized NumPy

Python Basics with NumPy

Practiced vectorized operations, broadcasting, and avoiding loops with NumPy to enable efficient deep learning computations.

Course: Deep Learning Specialization – Course 1, Week 2

💡 Foundation for Efficient Neural Network Computation

Reinforcement Learning

Applied Q-learning with function approximation to learn optimal policies in continuous state spaces using scalable policy and value function updates.

Course: ML Specialization – Course 3, Week 3

🤖 Q-learning in Continuous State Spaces

PCA and Data Visualization

Implemented Principal Component Analysis (PCA) for dimensionality reduction and data visualization, including eigen-decomposition, SVD, and high-dimensional data projection.

Course: ML Specialization – Course 3, Week 2

📉 PCA for Dimensionality Reduction and Visualization

Anomaly Detection

Implemented statistical anomaly detection using the multivariate Gaussian distribution to detect outliers in datasets through probability estimation and threshold tuning.

Course: ML Specialization – Course 3, Week 1

🔍 Gaussian-Based Outlier Detection from ML Specialization

Decision Trees

Trained a decision tree classifier on real-world data, visualized decision boundaries, and introduced ensemble methods like Random Forest and Gradient Boosting.

Course: ML Specialization – Course 2, Week 4

🌲 Implements Decision Trees and Boosting Techniques

Advice for Applying Machine Learning

Explored strategies for handling skewed data and evaluating model performance using precision, recall, F1-score, and error analysis.

Course: ML Specialization – Course 2, Week 3

⭐ From Andrew Ng’s ML Specialization – Course 2, Week 3

Neural Networks for Multiclass Classification

Built a neural network using softmax activation and cross-entropy loss for multiclass classification, with full vectorization for efficiency.

Course: ML Specialization – Course 2, Week 2

⭐ From Andrew Ng’s ML Specialization – Course 2, Week 2

Neural Networks for Binary Classification

Implemented a two-layer neural network using vectorized operations for binary classification with forward/backward propagation and visualization.

Course: ML Specialization – Course 2, Week 1

🧠 From Advanced Learning Algorithms – Week 1

Logistic Regression Implementation

Implemented logistic regression for binary classification using sigmoid activation and gradient descent, including L2 regularization to reduce overfitting.

Course: ML Specialization – Week 3

⭐ From Andrew Ng’s ML Specialization – Course 1, Week 3

Linear Regression Implementation

Implemented univariate linear regression using gradient descent to fit a line to data and visualize optimization with matplotlib.

Course: ML Specialization – Week 2

⭐ From Andrew Ng’s ML Specialization – Course 1, Week 2