I'm a Machine Learning Engineer with a strong passion for leveraging cutting-edge machine learning and deep learning techniques to solve complex business problems. My expertise lies in transforming raw data into actionable insights and building robust, end-to-end ML solutions with a strong focus on MLOps. I have a particular interest and niche for Computer Vision and Natural Language Processing (NLP).

I thrive on challenging projects that push the boundaries of what's possible with AI, and I'm committed to continuous learning and staying updated with the latest advancements in the field.

• 🔭 I’m currently developing an Agentic AI project and refining my end-to-end deep learning model for maize disease classification.
• 🌱 I’m currently deepening my knowledge in Large Language Models (LLMs), prompt engineering, and advanced NLP architectures.
• 👯 I’m looking to collaborate on complex MLOps initiatives and Agentic AI applications.
• 💬 Ask me about Machine Learning, Deep Learning, Data Science, Computer Vision, Natural Language Processing, MLOps, and TensorFlow.
• 📫 How to reach me: abdulrasheedolakiitan@gmail.com

Here are some of the tools and technologies I work with to build and deploy robust ML solutions:

Programming Languages:

Machine Learning & Deep Learning Frameworks:

Data Science & Analysis:

MLOps & Deployment:

Databases:

Version Control:

Here are some of the projects where I've applied deep learning, MLOps practices, and cutting-edge AI techniques to solve real-world problems. For more, check out my repositories!

This project tackles the critical challenge of identifying maize leaf diseases in natural environments, especially in Africa, where diseases like Maize Lethal Necrosis and Maize Streak Virus severely impact smallholder farmers' crop yields. My solution provides a pathway for early diagnosis to enhance food security.

• Dataset: Utilizes the largest publicly available dataset for maize leaf health classification, comprising 18,148 images captured with smartphone cameras in Tanzania. This dataset supports diverse tasks from disease diagnosis to real-time field prediction.
• Model Architecture: Implemented LFMNet, a lightweight multi-attention convolutional neural network. This architecture was specifically designed to overcome challenges such as background interference, high inter-class similarity, and to enable real-time inference in practical settings.
• MLOps Practices: Leveraged DVC for comprehensive data versioning and Docker for containerizing the model, ensuring reproducible and scalable deployment.
• Key Technologies: Python, TensorFlow, Keras, OpenCV, scikit-learn, DVC, Docker, Flask
• [Link to Repository (https://github.com/AbdulRasheed6/end-to-end_mazie_disease_classification)]

Developed an end-to-end AI-powered medical chatbot using Flask, LangChain, HuggingFace Transformers (all-MiniLM-L6-v2), Pinecone, and Gemini Pro to enable retrieval-augmented question answering over medical documents. The user interface was built with HTML and CSS while the backend handled semantic search and prompt-based response generation. The application was containerized with Docker and deployed to an AWS EC2 instance through a CI/CD pipeline using GitHub Actions, Amazon ECR, and a self-hosted runner. API keys and secrets were managed securely using GitHub Secrets, and model cache files were excluded via .gitignore.

• [Link to Repository ([(https://github.com/AbdulRasheed6/end-to-end_Medical_chatbot))]

Let's connect and discuss all things ML and Agentic AI!