Software engineer • Optimization researcher
I build reliable software—and algorithms for hard routing and scheduling problems.
I'm Michael, a full-stack engineer with an MSc in Computer Science. My work connects production-minded software engineering with evolutionary optimization and data-driven research.
Systems built for real constraints.
Full-stack products and operational tools—designed to be useful, testable, and maintainable.
Medical Equipment Rental Marketplace
A role-based rental platform connecting customers, equipment providers, and administrators through secure booking and inventory workflows.
Built REST APIs, JWT authentication, RBAC, Stripe payments, containerized services, and automated delivery workflows.
Point of Sale & Inventory System
Designed, deployed, and supported a POS and inventory platform, from requirements and implementation through testing and user training.
Two independent research works.
My research spans combinatorial optimization—solving real-world scheduling problems with evolutionary methods—and applied NLP, using transformer models to analyze low-resource language data.
Home Healthcare Routing & Scheduling
How do you assign skilled caregivers to patient visits across a city while simultaneously minimizing travel distance, waiting time, service delays, overtime, and workload imbalance—objectives that routinely conflict?
Build a decision-support system that surfaces trade-offs rather than hiding them, giving planners a structured set of Pareto-optimal schedules to choose from.
Multi-objective genetic algorithm with route-based encoding, Pareto ranking-based selection, and custom evolutionary operators tuned for the HHCRSP constraint structure.
Ghanaian Football Sentiment Analysis
Public sentiment around national football events is rich but noisy, multilingual, and culturally coded—standard NLP classifiers trained on Western corpora perform poorly on Ghanaian Twitter discourse.
Thousands of tweets collected around Ghana's 2022 FIFA World Cup campaign, covering match days, squad announcements, and post-tournament reactions.
Fine-tuned transformer models (BERT, RoBERTa) on a domain-specific labelled dataset built through targeted scraping and iterative annotation. Compared transformer performance against classical baselines.
Publicly released labelled dataset and classification benchmark for Ghanaian football discourse; demonstrates that domain-adapted transformers significantly outperform generic sentiment classifiers on this corpus.
Engineering depth.
I bring together software delivery, algorithm design, research, teaching, and support for the people who use the systems I build.
Web Developer/ Graphic Designer
Developing an AI-Enabled Knowledge Hub while maintaining and updating Future Black Female's WordPress and Wix websites, with a focus on usability, content management, accessibility, and reliable web experiences.
Backend Developer
Developed and supported backend and full-stack features using Python, Django, REST APIs, WebRTC, and Next.js, with a focus on reliability and maintainability.
Teaching & Research Assistant
Built and benchmarked Java optimization algorithms while teaching database systems, big-data programming, and media computation.
Software Developer
Designed, deployed, and supported a POS and inventory platform, from requirements and implementation through testing and user training.
EDUCATION
MSc, Computer Science
Brock University · 2025
BSc, Computer Science
KNUST · 2021
TOOLKIT