Building trustworthy AI systems through LLMs and RAG — interpretable, human-centered solutions for real-world applications. Backed by 4+ years of professional experience engineering secure web and data systems for international clients.
Building AI-powered systems that combine Large Language Models with interpretable rule engines for evaluating graduate programs, focusing on the intersection of machine learning accuracy and human-centered explainability.
Developed a neuro-symbolic RAG framework achieving 84% prediction accuracy while preserving full interpretability through deterministic rule-based reasoning.
Specializing in RAG, LLM fine-tuning, interpretable AI, and educational data mining. Certified in Generative AI (IBM, UMich) and a professional member of AAAI & ACM SIGAI.
Authored research demonstrating how fusing a deterministic symbolic engine with a neural semantic retriever mitigates LLM hallucinations and achieves 84% predictive accuracy.
Designed an interpretable AI system for evaluating 98 Brazilian graduate programs, combining LLM reasoning with symbolic scoring to achieve 84% accuracy with fully explainable outputs.
Comparative analysis framework for evaluating GPT-4o, Gemini 2.0, and DeepSeek across symbolic reasoning, prediction accuracy, and semantic tasks. Includes custom groundedness metrics and recall@k evaluation.
Gradio-based dashboard for real-time graduate program evaluation with multilingual support (EN/PT). Features cascading filters, HTML report generation, and integration with FAISS retrieval for document sourcing.
As my final year B.Sc. Project, I developed and deployed a centralized administrative platform to manage student data, significantly enhancing operational efficiency and record-keeping for a local educational institution in Nigeria.
Interpretable AI system combining LLM reasoning with symbolic scoring for analyzing graduate programs.
RAG-powered customer support chatbot for educational institution, reducing response time by 70%.
Interactive analytics dashboard for tracking KPIs with automated reporting and visualization.
Modern responsive website with integrated payment processing and inventory management for local business.
Designed a retrieval-augmented generation system linking policy documents with claim inquiries, focusing on embedding alignment for legal compliance.
Developed comparative visualizations of global pandemic data using R (ggplot2), communicating complex health trends to non-technical stakeholders.
Interested in research partnerships, PhD opportunities, or discussing AI applications in education.
Email MeAvailable for web development, AI integration, and data analysis consulting projects.
Connect with Me