Project Grapho
An end-to-end recommendation system using collaborative filtering and content-based algorithms.
Three areas I work across, from models to products that people use.
I build and evaluate AI models for prediction, classification, and personalization, from optimizing telemetry data to NLP classifiers and recommender systems, measured against real-world performance.
I design and ship systems built on large language models, like RAG systems, Agents, and evaluation harnesses. So generative AI works reliably inside real products and workflows.
I write production software around data, pipelines that move and transform it, APIs and dashboards that serve it — packaged in containers and deployed on cloud platforms.
Some projects I've worked on and contributed to
An end-to-end recommendation system using collaborative filtering and content-based algorithms.
An XGBoost classifier that predicts whether the flow around an aircraft is attached, near onset of separation, or separated, using simulated aerodynamic data.
An analysis of health facility records across Kaduna State, looking at where facilities are operational, registered, and licensed, and where the gaps are.
A service that turns raw sports commentary audio into polished 3–5 minute podcast episodes.
An NLP classifier that flags misleading news from writing style, reaching 99.21% test accuracy.
A voice-driven fashion retail demo where shoppers talk to an AI agent to browse products.
I lead the Colab ML community, where we build a variety of projects and research on foundations.
