Welcome to my Machine Learning Portfolio! I am Muhammad Abdullah, a Machine Learning Engineer passionate about leveraging data and cutting-edge technologies to solve complex problems. This repository serves as a showcase of my skills, experience, and the exciting projects I've worked on.
I bring a unique blend of technical expertise, project management, and collaboration skills to the table. With a focus on end-to-end machine learning solutions, I have successfully designed, deployed, and monitored models in real-world scenarios.
- Programming Languages: Python, Rust
- Tools and Frameworks: Scikit-learn, TensorFlow, PyTorch, Numpy, Pandas, Matplotlib, Seaborn, OpenCV, Fast API, Flask, Jupyter, Docker, Kubernetes, Terraform
- Big Data Technologies: Apache Spark, Hadoop
- Databases: MongoDB, SQLite, MySQL, PostgreSQL
- Cloud Platforms: AWS, GCP, Azure
- Soft Skills: Excellent communication, problem-solving, and collaboration
- Project Management:
- Tools: Trello, Notion
- Skills: Agile methodologies, project planning, execution
- The primary objective of this analysis is to enhance the fraud detection mechanism by refining the threshold used in the simulation. The emphasis lies in optimizing the criteria for identifying and capturing fraudulent transactions more effectively. Through a comprehensive Exploratory Data Analysis (EDA), we aim to derive valuable insights, patterns, and statistical summaries from the data, informing the formulation of an improved threshold strategy. This, in turn, contributes to the development of a more robust and accurate fraud detection system.
- Code: Link to Code
- Results: [Model Performance -> AUPRC = 0.9926360768575739]
- Visualisations: Visualization 1: Strip Plot
Visualization 2: 3D Scatter Plot
_Caption: A 3D Scatter Plot showing separation between Fraudulent and Non-Fraudulent Transactions using pre-existing and engineered features_Visualization 3: Diverging Palette
_Caption: A Diverging Palette to help visualise the difference in Footprint between Fraudulent and Non-Fraudulent Transactions._Explore my expertise in data analysis and cleaning through insightful Jupyter Notebooks and scripts. Gain insights into my EDA techniques and how I handle data cleaning challenges.
- Data Analysis: Link to Data Analysis Folder
- Data Cleaning: Link to Data Cleaning Folder
Explore my skills in MLOps, from version control to continuous deployment. See how I integrate machine learning into larger software projects and ensure smooth operations.
- MLOps: Link to MLOps Folder
- End-to-end Projects: Link to End-to-end Projects Folder
Check out projects deployed on various cloud platforms, showcasing my expertise in cloud computing, infrastructure as code, and scalable ML systems.
- AWS: Link to AWS Folder
- GCP: Link to GCP Folder
- Azure: Link to Azure Folder
Explore my proficiency in Big data, Generative AI, NLP and Text Analytics, Time Series Analysis, and Computer Vision. Also view my preferred sample ML Project Starter Code.
- Big Data: Link to Big Data Folder
- Generative AI: Link to Generative AI Folder
- NLP and Text Analytics: Link to NLP Folder
- Time Series Forecasting: Link to Time Series Forecasting Folder
- Computer Vision): Link to Computer Vision Folder
Connect with me on LinkedIn, Twitter, or drop me an email. I'm open to collaboration, discussions, and feedback.
- LinkedIn: Muhammad Abdullah
- Email: abdullahfast95@gmail.com
This repository is licensed under the [License Name]. Feel free to explore, contribute, and use the code in your own projects.