Courses
Advanced Learning from Data
Hardware Accelerator for Machine Learning
Machine Learning for Embedded Systems
AI Design and Deployment Risks
Human Robot Interaction
Big Data Technologies
Adversarial Machine Learning
Summary
As a Graduate Teaching Assistant, my responsibilities included leading weekly office hours for graduate students in an introductory machine learning course, evaluating a high volume of weekly assignments, and reinforcing core machine learning concepts. I provided targeted feedback on student projects, helping to enhance their practical application and understanding of tools like scikit-learn and TensorFlow.
Highlights
Led weekly office hours for 10+ graduate students, clarifying core machine learning concepts (such as supervised/unsupervised learning, regression, classification, and neural networks) and guiding their practical application in projects using scikit-learn and TensorFlow.
Evaluated 30+ weekly assignments for an introductory, graduate-level machine learning course, providing targeted, detailed feedback on student projects to enhance their practical application and understanding of ML principles with scikit-learn and TensorFlow.
Summary
Provided comprehensive software support and system development, improving operational efficiency and user engagement for national digital initiatives.
Highlights
Automated national address registry data collection using ODK Collect, Google Sheets, and Apps Script, achieving a 90% reduction in manual processing time.
Developed comprehensive workflow documentation to ensure seamless adoption and efficient data management practices.
Led the development of the Nigeria Egovernment Summit website (egovernment.ng) on WordPress, resulting in a 40% increase in user engagement and streamlined registration processes.
Provided critical IT support for over 200 attendees, ensuring smooth operations and successful event execution.
Customized a third-party digital sales platform using CSS to enhance user experience and functionality.
Significantly reduced support requests by developing clear documentation and comprehensive training videos, improving user self-sufficiency.
PyTorch, TensorFlow, Scikit-learn, Hugging Face Transformers, JAX.
Weights & Biases, MLflow, FastAPI, Gradio, Docker, Git, Amazon Sagemaker, Amazon Comprehend, Amazon Rekognition.
Python, C, SQL, Bash scripting.
PySpark, Polars, Pandas, Jupyter Notebooks.
HTML, CSS, Markdown, Wordpress.
Research
Summary
Tech Stack: PyTorch, Python ------------------------------------------ Developed a 128M-parameter GPT model from scratch in PyTorch, exploring core attention mechanisms and positional encoding. Optimized training for memory efficiency using gradient checkpointing and mixed precision, and experimented with sampling strategies like top-k and temperature to improve text coherence.
Academic
Summary
Tech Stack: PySpark, Geopandas, Jupyter Notebooks, Python ------- Conducted large-scale analysis of Chicago traffic congestion and collision patterns using PySpark and geospatial libraries, successfully identifying peak-hour hotspots and high-risk zones. Key insights were visualized within Jupyter Notebooks.
Summary
Tech Stack: Hugging Face Transformers, PyTorch, FastAPI, Gradio, Python Summary: Implemented a customer feedback classification system utilizing a pretrained MNLI transformer model (via Hugging Face Transformers and PyTorch). Deployed this system as an interactive API using FastAPI and Gradio, enabling real-time inference.
Academic
Summary
Designed and trained a Deep Convolutional GAN (DCGAN) to generate solvable mazes with optimized complexity. Adversarial training techniques were leveraged to significantly enhance the realism and diversity of the generated mazes.
ECE Department
National Society of Black Engineers (NSBE) GMU Chapter.