Sharon Muiruri profile picture

Hello, I'm

Sharon Muiruri

Data Scientist | Machine Learning Engineer

My Github Profile My Linkedin Profile

Get to know more

About Me

An ambitious, data-driven Entry Level Data Scientist eager to leverage analytical and technical skills to identify and solve complex challenges, create meaningful insights, and drive business value. With a proven track record of delivering innovative solutions and improving data efficiency, I am committed to developing and deploying successful data platforms that meet the business's needs. I am passionate about gaining new skills and I aim to excel in my career through continuously learning and seeking opportunities that widen my capabilities.




My Skills

Skills

Programming Languages: Python
Machine Learning & Deep Learning:Scikit-learn, TensorFlow, PyTorch, Langchain; expertise in Regression, Classification, Clustering, NLP, Computer Vision, Neural Networks, LLMs
MLOps:Docker, Kubernetes
Databases: SQL, MongoDB
Cloud Platforms: Microsoft Azure
Version Control: Git
Arrow Icon

Explore my

Experience

Experience icon

Work

Sep 2024 - present

Machine Learning Engineer Intern

Codsoft

- Engineered data pipelines and developed predictive machine learning models to solve complex business challenges.
- Built and deployed a customer churn prediction model, leading to increased retention rates through targeted interventions.
- Enhanced the accuracy of a fraud detection classification model by conducting thorough data preprocessing and feature engineering.

Sep 2022 - Dec 2022

Data Science/MLOps Intern

Flapmax AI Institute

- Implemented and customized a Commerce Marketplace SaaS Accelerator on Microsoft Azure, enhancing scalability and performance.
- Deployed two FastAPI applications using Kubernetes and Docker for efficient containerization and orchestration.
- Collaborated with cross-functional teams to develop a car damage detection model using Mask-RCNN, improving object detection accuracy.

Education icon

Education

2022-2023

Data Science and Machine Learning

Massachusetts Institute of Technology

Grade: A

- Enrolled in Data Science and Machine Learning program where I gained skills in python, statistics, data analysis and visualization, machine learning, deep learning and computer vision.

2018-2022

Business Information Technology

Jomo Keyatta University of Agriculture and Technology

Grade: Second Class Honours

- Pursued a Bachelor's degree in Business Information Technology where I acquired skills in various areas such as programming, data structures and algorithms, networking and database systems.

Arrow Icon

Browse My Recent

Projects

Stock Market Prediction

Stock Market Prediction

  • The objective was to develop a model to forecast stock market prices.
  • I began with data collection, scraping stock data from the yfinance API library, followed by conducting exploratory data analysis to explore trends, volatility, and relationships between stock prices.
  • To improve the model performance, I applied feature engineering, using lag variables to capture dependencies in stock prices across different time periods.
  • I built and evaluated 3 models: ARIMA, SARIMA, and LSTM. The LSTM model achieved the best performance, demonstrating the lowest RMSE score, indicating its ability to effectively predict stock prices.
Credit Card Default Prediction

Credit Card Default Prediction

  • I developed a predictive model aimed at identifying customers likely to default on credit card payments, enabling the company to mitigate financial risks and enhance credit decision-making.
  • I cleaned, analyzed and preprocessed the data, then implemented SMOTE to balance the dataset and ensure robust model performance.
  • Built and evaluated five classification models: Logistic Regression, Decision Trees, Random Forest, XGBoost and Gradient Boosting, with Random Forest achieving the highest ROC AUC score.
  • The model not only improved the company's ability to proactively manage credit risk but also enhanced its ability to tailor customer outreach and optimize collection strategies.
Customer Reviews Analysis

Customer Reviews Analysis

  • The aim of the project was to analyze reviews provided by British Airways customers and present insights that would influence decisions and make a tangible impact on the business.
  • I scraped the data from Skytrax then cleaned it, and conducted extensive analysis to understand the customers' feelings, needs and feedback.
  • Performed sentiment analysis and built a wordcloud to provide some insight into the content of the reviews.
Store Sales Forecasting

Store Sales Forecasting

  • I developed a predictive model to forecast six weeks of daily sales for 1,115 Rossmann stores, providing the company with actionable insights for inventory management and sales optimization.
  • Leveraging Python libraries such as scikit-learn and pandas, I conducted comprehensive data analysis and modeling.
  • The data was preprocessed by imputing missing values and addressing outliers through Z-score treatment to enhance model accuracy.
  • I built and evaluated 4 regression models: Linear Regression, Stochastic Gradient Descent, Decision Trees, and Random Forest.The Random Forest model delivered the best performance
Hotel Booking Cancellation Prediction

Hotel Booking Cancellation Prediction

  • The objective of this project was to develop a predictive model to identify hotel bookings likely to be canceled in advance, enabling the formulation of more effective and profitable cancellation and refund policies.
  • I performed univariate and bivariate analysis to identify key factors influencing booking cancellations.
  • Built and evaluated 4 classification models: Logistic Regression, Support Vector Machine, Decision Trees and Random Forest.The best performing model was the Random Forest with an accuracy score of 90%.
Arrow Icon

Get In Touch

Contact Me

Phone icon

Call Me

+254-745-728-616
Email icon

Email

sharnmuiruri44@gmail.com

Location icon

Location

Nairobi - Kenya