Abiola Onasanya
Software Engineer building backend systems and machine learning applications.
MSc Artificial Intelligence, Sheffield Hallam University. Based in the UK.
Formerly Software Engineer at Validators Innovation.
Selected work
Pediatric Pneumonia Classification
MSc Dissertation · 2026
MSc dissertation measuring whether lightweight pediatric pneumonia classifiers transfer asymmetrically between Chinese and Nigerian chest X-ray cohorts, with a reproducible confidence-reliability evaluation framework.
FIG-Loneliness
AI Research and Development Project · Team project · 2026
A team research project on automated loneliness self-disclosure detection in Reddit posts, where I built the deployed FastAPI backend, the inference engine, and the eight-page interactive dashboard that made the ML pipeline usable.
StoryReasoning
Deep Neural Networks and Learning Systems · 2025
A modular, lightweight toy implementation inspired by the StoryReasoning paper, exploring grounded visual storytelling with chain-of-thought grounding tables at small scale.
ZSPMS
Advanced Robotics and Artificial Intelligence · 2025
A Webots-based autonomous package-sorting system using a UR10e arm, custom-trained YOLOv8 detection, OCR and CLIP verification, and a joint-sensor-driven finite state machine.
Experience
Work
Aug 2022 – Aug 2025
Software Engineer · Validators Innovation, Lagos, Nigeria
Designed and implemented backend software systems and RESTful APIs using Python, Django and Django REST Framework within a logistics technology environment.
Dec 2021 – Aug 2022
Junior Software Engineer · Fastrade Logistics Limited, Lagos, Nigeria
Contributed to software development activities across backend and application functionality.
Education
2025 – 2026
MSc Artificial Intelligence · Sheffield Hallam University, United Kingdom
Dissertation on the statistical evaluation of cross-population generalisation and calibration in chest X-ray classifiers.
2019 – 2022
BSc Computer Science · ESAE University, Benin Republic
Taught primarily in English, with selected modules delivered in French.