Opeyemi Adeniran

Research: Early Detection of Alzheimer's Disease Using an Ensemble of Diverse Convolutional Neural Networks and Vision Transformer.

• Implemented array of preprocessing techniques, including: Normalization, Augmentation Transformation to Tensorflow and Numpy formats on 70,706 MRI images
• Designed and trained seven state-of-the-art models: DenseNet201, EfficientNetB0, MobileNetV2, ResNet50, VGG19, Vision Transformer (ViT), and custom model named "Multi_classifier_CNN". Achieved individual model accuracy ranging from 46.3% to 98.0% for classifying patients into three distinct classes: Alzheimer's disease, mild cognitive decline, and cognitive normal.
• Integrated Grad-CAM and LIME visualization techniques were used to interpret and highlight critical regions in MRI images, which influenced model predictions. Categorized patients into three distinct classes: Alzheimer's disease, mild cognitive decline, and cognitive normal.
• Leveraged ensemble techniques, specifically "hard voting" and "stacking,", to refine diagnosis accuracy. Achieved commendable accuracy of 99.0% with hard voting and 99.33% using stacking, emphasizing model's effectiveness and reliability.

• Formulated and executed SQL queries to facilitate seamless migration of exiting excel worksheet to AWS cloud environment, ensuring data integrity, availability, and minimal downtime.
• Actively collaborated with IT team in diagnosing and resolving challenges related to cloud data migration and optimizing web page load times, resulting in 98% improved system reliability and user experience.

• Developed experiment using experiment-centric pedagogy teaching approach in 3 consecutive semesters to teach 30 computer science students taking digital logic course for with focus on multiplexers, full adders, subtracter, and fundamental logic gates to encourage, motivate, and improve students learning outcomes and performance in class.
• Performed statistical, qualitative, and quantitative analysis.
• Wrote and developed research papers with focus on use of experiment-centric teaching approach to enhance students learning performance.

• Developed and deployed Chatbot leveraging Amazon Lex on SLACK to enlightening children about online child abuse risks. Natural Language Processing (NLP) was used to analyze data from social media platforms and utilized Word cloud for visual representation of prominent topics.
• Led team of 40 machine learning engineers to develop models that identified top Nigerian regions with high demand for electricity and pinpoint optimal locations for solar panel installations. Extracted 774 nighttime raster images (Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB) (2012–2021) from Google Earth Engine (GEE), layered LGA shapes files on VIIRS/DNB, and calculated mean value of each zone using zonal statistics. Population dataset for each LGA was extracted from WorldPop and merged with average nighttime mean. Applied K-means clustering, which categorized areas into five distinct energy demand categories: low, moderately low, medium, moderately high, and high intensity. Final model was deployed using Streamlit app, providing stakeholders with clear and interactive overview of areas with population along with their energy demand.
• Led team of 20 machine learning engineers to develop recommendation system that suggest job opportunities to graduates based on people's profiles. 120 public datasets was scraped from LinkedIn using Beautiful Soup library, TF-IDF (Term Frequency Inverse Document Frequency) algorithm was used to transform text data into numerical data, and finally Adopted Word cloud for data visualization. recommendation model was built on content-based filtering, leveraging both cosine similarities and nearest neighbor algorithms for precise job matching.
• Got promoted as AI Evangelist, and secured access to free learning of about $500,000 whats of scholarships. Spoke on international platforms such as TechSoup and among local communities.

• Underwent comprehensive training in Data Science, acquiring proficiency in SQL, Excel, Tableau, and Python libraries for data analysis and model development
• Developed comprehensive Tableau dashboard by seamlessly integrating data from "Assignment" and "Leaves" files, providing managers with real-time insights into project status and team availability. This Empowered managers to have capability to track and oversee their specific projects; anticipate and plan around team member holidays up to 3 months in advance; make informed decisions regarding holiday approvals based on current leave schedules; differentiate between active (ongoing) and inactive (upcoming or completed) projects; access specific monthly views, with special focus on months like September and October.
• Implemented visual alerts to flag projects with potential risks, especially those with over 50% of team members on scheduled leave, ensuring proactive managerial interventions.
• Introduced tailored filters, including company-specific (Astek and Roche) and manager-specific views, ensuring personalized user experiences and enhanced decision-making capabilities.

• Imparted essential web design principles to more than 200 students using core languages such as HTML, CSS, and JavaScript
• Delivered comprehensive instruction in foundational programming languages, including C, C++, and C# to more than 200 students
• Equipped students with skills necessary to write, test, and refine code,