• Native iOS stock market analysis app using SwiftUI, Xcode, and Alamofire, integrating Finnhub API for real-time data and virtual trading features. Implemented portfolio management, favorites tracking, and interactive data visualization with a Node.js backend and MongoDB for data persistence. [Github Link]

  *

• Designed and executed a project to optimize skiing routes to ski resorts, incorporating advanced pathfinding techniques such as BFS, UCS, and A* search algorithms. [GitHub Link]
• Devised min-max algorithms permitting a bot to accurately identify and execute capture moves based on game rules for Pente, resulting in a 20% improvement in the AI's competitiveness and strategic thinking ability in gameplay. [GitHub Link]
• Developed and implemented a robust first-order logic resolution system to utilize a knowledge base encoding all restaurant details, enabling the program to generate logical conclusions with accuracy and precision. [GitHub Link]

*

• A Binary Classifier of vertebral column dataset using K-Nearest-Neighbors. [GitHub Link]
• A Linear, Multiple, and KNN Regressor to predict the Electrical Energy Output of Combined Cycle Power Plant. [GitHub Link]
• A classifier of human activities based on time series obtained by a WirelessSensor Network. [GitHub Link]
• A decision tree to determine if a patient has accute inflammation or has accute nephritis based on bodily symptoms. [GitHub Link]
• A Lasso and L1 penalized gradient boosting tree for communities and crime dataset. [GitHub Link]
• A Random Forest And XGBoost Model Trees for component failure detection in the APS system. [GitHub Link]
• A Multi-class and Multi-Label Classification Using Support Vector Machines, K-Means Clustering, and Monte Carlo Simulation for species, genus and family detection of frogs. [GitHub Link]
• Supervised, semi-supervised, and unsupervised learning methods, designed and evaluated Monte Carlo simulations, active learning strategies, and spectral clustering techniques to analyze performance metrics [GitHub Link]
• Designed and implemented a multi-class image classifier for 20 bird species using transfer learning with EfficientNetB0 and VGG16, incorporating data augmentation techniques and fine-tuning the final layers. [GitHub Link]

*

• Developed a sentiment analysis pipeline on the Amazon Reviews dataset using Python and libraries like Scikit-learn and NLTK. Implemented data preprocessing, TF-IDF vectorization, and machine learning models (Perceptron, SVM, Logistic Regression, Naive Bayes) to classify text. Achieved strong performance metrics (accuracy, precision, recall, F1-score) and demonstrated expertise in text classification, feature engineering, and model evaluation. [GitHub Link]
• Developed a sentiment analysis pipeline using Word2Vec embeddings (pretrained and custom-trained) for feature extraction, combined with machine learning models (Perceptron, SVM) and deep learning architectures (MLP, CNN) for binary and ternary text classification. Conducted comparative evaluations using TF-IDF, pretrained, and custom Word2Vec features, demonstrating expertise in feature representation, model training, and performance analysis. [GitHub Link]
• Developed a Hidden Markov Model (HMM) for Part-of-Speech (POS) tagging using the Penn Treebank dataset, incorporating Greedy and Viterbi decoding algorithms for sequence prediction. Implemented vocabulary handling, optimized emission and transition probabilities, and evaluated accuracy using custom evaluation scripts, automating result generation for development and test datasets. [GitHub Link]
• Developed a Named Entity Recognition (NER) system using Bidirectional LSTM with GloVe embeddings, further optimized with an LSTM-CNN architecture. Achieved high precision, recall, and F1 scores on the CoNLL-2003 dataset, showcasing expertise in deep learning and NLP. [GitHub Link]

*