Favour tumba

I’m a Computer Science graduate with First Class Honours (GPA 4.57/5.00), focused on building AI and backend systems in Python. I specialize in turning LLM capabilities into production-ready applications—fast, measurable, and reliable.

On my LLM Evaluation & Observability Platform, I trace every LLM call (prompt, response, tokens, latency, and cost) through a FastAPI ingestion API backed by PostgreSQL, and I provide an SDK plus a Streamlit dashboard to monitor cost, latency, and error trends. On my Document Intelligence API, I ingest PDFs, chunk and embed them, and deliver answers with structured citations using the Claude API, combining PostgreSQL full-text search with pgvector embeddings via Reciprocal Rank Fusion.

I’ve also built end-to-end ML workflows in Python, including a leakage-free scikit-learn pipeline and a Streamlit app for Heart Disease Risk Prediction (SMOTE, ROC-AUC evaluation). As an IT Intern at NITDA, I automated and streamlined operational data handling—improving data management efficiency and team accessibility—while collaborating to reduce redundant work and support digital literacy initiatives.