The Engineer Behind the Code
I build ML systems that work on real, messy data — not clean benchmarks. Computer Science graduate from Landmark University. My thesis achieved 87.1% ROC-AUC on multimodal breast cancer detection — a result driven by architectural decisions, not luck.
My Story
I'm a Computer Science graduate from Landmark University with a sharp focus on applied machine learning and data engineering — specifically, making AI work outside controlled environments.
My thesis was the clearest demonstration of this. Rather than train another single-modality classifier on a clean benchmark, I designed a multimodal fusion system combining ultrasound image embeddings (MedCLIP) with clinical and radiometric tabular data (TabNet). Result: 87.1% ROC-AUC — significantly outperforming unimodal baselines.
At Kontemporary Konsulting Ltd (Aug–Sep 2025), I architected a Selenium-based scraping system navigating multi-layer dynamically rendered structures across all 36 Nigerian states, loading everything into a clean queryable PostgreSQL database. 99.9% accuracy. 90% time reduction.
At TETFund (Mar–Sep 2024), I replaced manual report processing workflows with Python automation pipelines — handling messy institutional Excel files, correcting inconsistencies, and migrating legacy records into structured databases.
I write code that solves specific real problems. I document what I build. I work well with imperfect data — because that's all there is in the real world.
Experience Timeline
87.1% ROC-AUC via MedCLIP + TabNet early fusion architecture.
Built Selenium electoral data infrastructure across 36 states (99.9% accuracy). Benchmarked ViT vs DINOv2 on brain tumor MRI with Grad-CAM visualization.
Automated report processing pipelines. Migrated legacy Excel records to PostgreSQL with full validation.
Algorithms, data structures, machine learning, software engineering. Final year thesis on multimodal AI for medical diagnostics.