Michael Yemenshwa

High-performing computer scientist graduate, officially transfering to become a Terp at the prestigious University of Maryland, with a passion for coding and experience in troubleshooting Linux technical issues and handling big datasets expressed in my GitHub. Seeking a challenging tech internship to apply my skills and gain further experience in the industry, providing excellent research, technical and problem-solving skills with a detail-oriented mind, aiding me in being able to learn new concepts quickly.

• PythonProjectsRepo(filename: KNN2.py): Conducted in-depth Exploratory Data Analysis (EDA) on the Iris dataset using Pandas and Seaborn, highlighting separability among species through pair plots., Demonstrated proficiency in data visualization by creating Kernel Density Estimation (KDE) plots to analyze distribution of key iris attributes., Employed scikit-learn for robust data preprocessing, including feature standardization using StandardScaler, and split data into 70% training and 30% testing sets., Applied K-Nearest Neighbors (KNN) classification algorithm to accurately predict Iris species, achieving an initial accuracy of 96%., Employed data-driven decision-making to optimize model performance, systematically exploring error rates for different K values and retraining the model for a 100% accuracy rate with the best K value.
• ReinforcementLearning(filename: QLearning.py): Engineered a Q-learning algorithm to solve a navigation problem, leveraging matrices to guide optimal decision-making., Conducted comprehensive training over 10,000 iterations, enabling the agent to learn actions and build an efficient policy., Demonstrated advanced algorithmic skills by successfully implementing and interpreting the Q-learning process, showcasing expertise in reinforcement learning and problem-solving.