Results-driven Machine Learning Engineer with hands-on expertise in Computer Vision, Natural Language Processing, and Conversational AI. I specialize in designing and deploying end-to-end machine learning solutions that achieve 76–98% model accuracy across real-world applications.
Recently completed a 6-month industrial training at University College Hospital (UCH), Ibadan, where I developed AI tools for healthcare analytics while working with privacy-sensitive medical data.
Currently pursuing: B.Tech. in Computer Science at Ladoke Akintola University of Technology (LAUTECH) | Expected 2026 | GPA: 4.3/5.0
Programming Languages: Python, SQL
Frameworks & Libraries: TensorFlow, Keras, PyTorch, Scikit-learn, Pandas, NumPy
AI/ML Domains: Computer Vision, NLP, Large Language Models (LLMs), Conversational AI
Deep Learning: CNNs, Transfer Learning, Embeddings, Sequence Modeling
Tools & Platforms: Streamlit, Git/GitHub, Jupyter, Google Colab, Kaggle
Data Science: Data Wrangling, Model Evaluation, Matplotlib, Seaborn
Tech Stack: AutoGen • LLM • Streamlit • Groq API
Production-ready medical chatbot delivering safe, empathetic, multi-turn conversations with patients.
- Integrated Microsoft AutoGen Framework for agentic reasoning and contextual dialogue management
- Implemented emergency-response detection and medical disclaimer protocols for responsible AI behavior
- Deployed Streamlit-based web app with chat history, quick-actions, and intuitive user interface
- Ensures HIPAA-aware conversation flow with safety filters
Tech Stack: NLP • Universal Sentence Encoder • Conv1D • TensorFlow
Multi-label classification system for automating biomedical literature review.
- Achieved 80% accuracy on PubMed abstract classification across multiple categories
- Engineered hybrid architecture combining USE embeddings with Conv1D for sequential reasoning
- Processed 200K+ scientific sentences, dramatically reducing manual literature triage time
- Enables researchers to rapidly identify relevant studies by abstract section
Tech Stack: Computer Vision • MobileNet • Transfer Learning • TensorFlow
Deep learning classifier for fine-grained dog breed recognition.
- Trained CNN model achieving 76% validation accuracy across 120 dog breeds
- Evaluated multiple architectures, selecting MobileNet for optimal performance-to-efficiency ratio
- Built complete ML pipeline: data preprocessing → augmentation → training → inference
- Implemented real-time prediction interface for user uploads
Tech Stack: EfficientNetB0 • Transfer Learning • TensorFlow • Keras
Fine-grained classification system for the CARS196 dataset.
- Fine-tuned EfficientNetB0 on 16,000+ images across 196 car model classes
- Performance: 99% training accuracy | 78% validation accuracy | 75% test accuracy
- Implemented advanced data augmentation strategies to combat overfitting
- Deployed on Google Colab for accessible experimentation
Tech Stack: CNN • TensorFlow • Keras
Binary image classification achieving exceptional performance.
- Built custom Convolutional Neural Network from scratch
- Achieved 98%+ accuracy on 11,000+ test images from Kaggle
- Trained on FreeCodeCamp's Cat & Dog Challenge Dataset
- Demonstrates mastery of fundamental computer vision techniques
Tech Stack: Pandas • Scikit-learn • XGBoost
Binary classification model for identifying potential mental health concerns from survey data.
- Performed comprehensive EDA and feature engineering on sensitive healthcare data
- Implemented multiple ML algorithms with hyperparameter tuning
- Delivered actionable insights for early intervention strategies
Tech Stack: Random Forest • Scikit-learn • NumPy
Automated loan approval decision system for financial institutions.
- Developed Random Forest Classifier achieving 90% accuracy
- Engineered features from applicant financial and demographic data
- Created interpretable model supporting fair lending practices
Tech Stack: LightGBM • Pandas • Seaborn
Regression model predicting user engagement metrics.
- Built LightGBM regressor on custom podcast listening dataset
- Performed feature engineering from temporal and behavioral data
- Delivered insights for content optimization and user retention
Industrial Trainee – Machine Learning / AI
University College Hospital (UCH), Ibadan | April 2025 – October 2025
- Completed 6-month industrial training applying AI solutions to medical workflows
- Developed and evaluated ML models for healthcare data analysis with strict privacy standards
- Collaborated with multidisciplinary teams to deploy intelligent systems supporting hospital operations
- 🎓 Machine Learning with Python – freeCodeCamp (2024)
- 🎓 Deep Learning with PyTorch – Udemy (2024)
- 🎓 Google Data Analytics Professional Certificate – Coursera (2024)
✅ Deployed 3+ production-ready ML applications with live web demonstrations
✅ Delivered 76–98% accuracy across Computer Vision and NLP tasks through optimized pipelines
✅ Active Kaggle contributor with strong competition performance
✅ Built comprehensive AI-focused portfolio showcasing end-to-end project lifecycle
- 📧 Email: amusanolanrewaju420@gmail.com
- 💼 LinkedIn: linkedin.com/in/olanrewaju-amusan
- 🌐 Portfolio: st-lexy.github.io
- 📱 Phone: +234 704 577 3561
💡 Open to collaboration on AI/ML projects, research opportunities, and challenging problem-solving initiatives.
⭐ If you find my projects useful, consider giving them a star!