Sadik Bulut

AI & ML Engineer | Data Scientist | Solutions Architect

About Me

Hello👋 I'm Sadik Bulut, a Data Scientist and certified AWS + Azure Solutions Architect with advanced AI & ML training, and a proven ability to leverage data analytics and visualization to solve complex problems.

Projects

Project D: Brain Tumor Image Classifier

Keywords: Convolutional Neural Networks, VGG-16 CNN, Deep Learning, Image Classification, TensorFlow, Keras, Data Augmentation, Transfer Learning, Medical Imaging

This project builds a binary image classifier to distinguish Pituitary Tumor in MRI scans of patients using a Kaggle dataset. A subset of 1,000 images is used, with data augmentation to prevent overfitting. Transfer Learning with a VGG-16 pre-trained model is applied to enhance performance, showcasing robust medical image classification techniques.

Click HereBrain Tumor Classifier

Project C: Bank Customers Churn Prediction

Keywords: Neural Networks, Deep Learning, Customer Churn Prediction, TensorFlow, Keras, SMOTE, Class Weight, Customer Relationship Management (CRM), Predictive Analytics

This project leverages customer data analysis to identify key factors that influence customer retention and utilizes this insight to develop a Neural Network Model. The designed model aims to predict the likelihood of a customer churning within the next 6 months. By harnessing this predictive capability, the model can be effectively used in customer relationship management strategies to enhance retention rates. Applications include proactive customer engagement, personalized offers, and targeted support interventions, all crucial for reducing churn and improving service satisfaction.

Click HereBank Churners Prediction

Project B: Audio Digit Recognition

Keywords: Artificial Neural Networks, Librosa, TensorFlow, Keras, Deep Learning, Audio Processing, Speech Recognition

This project utilizes the Librosa library to visualize the Audio MNIST digit dataset and converts it into a numerical database for training an Artificial Neural Network (ANN). The trained model is designed for speech recognition applications, specifically for recognizing spoken digits. Such a digit speech recognition model can be applied in various domains, including voice-activated security systems using digit-based codes, automated telephone operators, and similar use cases.

Click HereAudio Recognition

Project A: Conversational Chatbot Powered by OpenAI API

Keywords: OpenAI API, Generative AI, Retrieval-Augmented Generation, JavaScript, AWS Amplify

This ChatBot utilizes OpenAI's API to integrate the ChatGPT 3.5 language model, enabling interactive conversations with users. The project is developed using JavaScript and deployed on AWS using Amplify for seamless scalability and hosting.

Click HereOpenAI API ChatBot

Project 9: Credit Card Users Churn Prediction

Keywords: Machine Learning, GradientBoosting, GXBoost, Logistic Regression, SMOTE, Cross Validation, Churn Prediction, Hyperparameter Tuning

This project is my favorite so far! It evaluates 18 machine learning models to predict whether a bank's credit card users will churn or continue using the service. After identifying the top three models, it fine-tunes their hyperparameters to determine the best-performing model for this type of data.

Click HereCC Users Churn Prediction

Project 8: Employee Attrition Estimator

Keywords: Random Forest, Bagging, Machine Learning, Model Evaluation, HR Analytics, Talent Retention

This Employee Attrition Estimator leverages a Random Forest machine learning model to predict employee turnover, identifying key factors contributing to attrition. This tool helps organizations proactively retain talent by providing insights into which employees are at risk of leaving.

Click HereEmployee Attrition Estimator

Project 7: Decision Tree Classifier: Personal Loan Campaign

Keywords: EDA, scikit-learn, Model Building: Decision Tree, Pruning, Supervised Learning, Model Performance Evaluation & Improvement

This machine learning project uses a decision tree model to predict whether liability customers at AllLife Bank will convert to personal loan customers, helping the bank identify high-potential customers for targeted marketing and personalized loan offers.

Click HereLoan Conversion Prediction

Project 6: Webpage Scraper

Keywords: Web Scraping, BeautifulSoup, Data Extraction, Treemap, Squarify

A simple web page scraper using BeautifulSoup and Requests libraries to pull data from a Wikipedia page. This data is then visualized with Squarify to create a Treemap, illustrating the revenues of top companies categorized by industry.

Click HereWeb Scraper

Project 5: Education vs. Economy

Keywords: Interactive Visualizations, Matplolib, Seaborn, Plotly, statsmodels

Animated graphs explore the relationship between years of schooling and GDP per capita across countries over time. Colored by country characteristics, the plots include a trendline to highlight correlations, providing insights into education's impact on economic growth.

Click HereEducation vs Economy

Project 4: COVID-19 Detection from X-Rays

Keywords: Image Processing, Python, OpenCV, TensorFlow, Keras, Seaborn, Gradio

A COVID-19 detection project utilizing X-ray images aims to identify signs of infection through advanced image analysis. The model enhances diagnostic accuracy, aiding healthcare professionals in timely decision-making, and improving patient outcomes while leveraging AI for efficient image interpretation.

Click HereCOVID-19 Detection

Project 3: FoodHub Order Analysis

Keywords: Python, NumPy, Pandas, Matplotlib, Google CoLab, Exploratory Data Analysis (EDA), Univariate & Bivariate Analysis, Business Recommendations

An exploratory data analysis (EDA) project on FoodHub’s order data, revealing trends, optimizing delivery times, and enhancing customer satisfaction. Insights include popular cuisines, customer feedback, and revenue impact. Recommendations focus on improving ratings, promoting top restaurants, and streamlining delivery operations.

Click HereFoodHub Order Analysis

Project 2: Atom Simple

Keywords: Next.js, React, JavaScript, DynamoDB, AWS Cognito, AWS Amplify

A web application utilizing Amazon Cognito for user authentication, enabling sign-up and sign-in functionality. Registered users can access and interact with dynamic storybooks. User metadata is securely stored in a DynamoDB database. The entire application is deployed and managed through AWS Amplify for seamless scalability and integration.

Click HereAtom Simple

Project 1: AI TutorBot

Keywords: AWS Bedrock, Generative AI, Knowledge base, Conversational AI, LLM

Developed an intelligent, generative AI-powered chatbot using AWS Bedrock and foundation models for customer support. The chatbot leverages uploaded documents to build a comprehensive knowledge base, enabling natural, conversational interactions. It can be seamlessly integrated into a website to engage with customers and provide personalized assistance.

Click HereAI TutorBot

Project 0: Visualizing Canvas Discussion Interactions

Keywords: Social Network Analysis (SNA), Socail Network Diagram (SND), Discussion Analytics, Interaction Mapping, Web Scraping

These graphs scrape data from discussion forums in a Canvas course, creating Social Network Diagrams (SND) to make Social Network Analysis (SNA). They visualize interaction patterns among participants, highlighting key contributors and engagement levels to enhance understanding of collaborative learning dynamics.

Click HereSocial Interaction Mapping