Back
PORTFOLIO / PROJECTS

My Projects

A curated collection of projects showcasing my expertise in web development, design systems, and modern technologies.

Contributions
Less
More

Mineral Deposit Prospectivity

Machine Learning Engineer & Geospatial Analyst

January 2026 – Present

An ongoing machine learning project focused on predicting mineral deposit locations using advanced geospatial analysis, geological data, and AI models. The system identifies high-prospectivity areas for mineral exploration, optimizing resource discovery and reducing exploration costs.

  • Geospatial Data Integration: Combines multiple geological datasets including geological maps, geophysical surveys, geochemical data, and remote sensing imagery for comprehensive analysis
  • Machine Learning Models: Develops predictive models using ensemble methods and deep learning to identify patterns indicative of mineral deposits
  • Prospectivity Mapping: Generates high-resolution prospectivity maps highlighting areas with highest probability of containing mineral deposits
  • Multi-Source Data Fusion: Integrates diverse data sources including satellite imagery, geological surveys, and historical exploration data for enhanced accuracy
  • Risk Assessment: Provides confidence metrics and uncertainty quantification to support exploration decision-making
  • Cost Optimization: Aims to reduce exploration costs by prioritizing high-prospectivity areas and minimizing unnecessary drilling and sampling
TECHNOLOGIES
Python Machine Learning Geospatial Analysis Remote Sensing GIS Deep Learning Data Fusion Geological Modeling

Clothing Image to Sewing Pattern & 3D Model

Machine Learning Engineer & 3D Graphics Developer

December 2025 – Present

An advanced machine learning system that analyzes clothing images and generates accurate sewing patterns along with physics-based 3D models. The system extracts garment structure from photographs and creates both 2D pattern templates and realistic 3D simulations with material physics.

  • Image Analysis & Pattern Extraction: Uses computer vision and deep learning to analyze clothing images and extract garment structure, seams, and construction details to generate accurate sewing patterns
  • Sewing Pattern Generation: Automatically generates 2D sewing pattern templates with measurements, seam allowances, and construction guidelines based on the analyzed garment
  • 3D Model Reconstruction: Creates detailed 3D models of the clothing with accurate geometry, topology, and surface details matching the original garment
  • Physics-Based Material Simulation: Implements realistic material physics including fabric drape, stretch, weight, and texture properties for accurate 3D representation
  • Material Property Detection: Analyzes fabric type, texture, and properties from images to apply appropriate physics parameters in the 3D model
  • Interactive 3D Visualization: Provides interactive 3D models with working physics simulation, allowing users to see how the garment behaves under different conditions
TECHNOLOGIES
Python Machine Learning Computer Vision Deep Learning 3D Modeling Physics Simulation Image Processing Pattern Recognition Fabric Simulation

Ananse x Visio

Co-Founder

2025 – Present

A collaboration between Ananse and Visio bringing education to those far away. An offline AI-driven educational system designed to bridge the education quality gap in Ghana's rural regions without relying on the Internet.

  • Operates completely offline with entire GES-aligned syllabus stored locally using SQLite and ChromaDB
  • No internet connection required at any stage of the learning process
  • Pre-generated and AI-generated content blend seamlessly to enhance learning
  • All progress and interaction data stored locally to ensure privacy and safety
  • USSD-based support system for repairs and system replacements in remote areas
  • Deployment timeline: MVP (June 2025), Refinement (October 2025), Full Deployment (December 2025)
TECHNOLOGIES
AI/ML Offline Systems SQLite ChromaDB USSD Education

AfriArt - NFT Art Platform

Developer - 1st Place Winner

2025

Platform designed to empower student artists to showcase, auction, and sell their artwork as NFTs. Built on Internet Computer Protocol (ICP) using Motoko. Winner of 1st Place at Akwaaba Dappathon (ICP HUB Ghana).

  • Developed NFT marketplace platform enabling student artists to monetize their artwork
  • Implemented auction system for artwork sales with blockchain-based transactions
  • Built on ICP blockchain using Motoko programming language for smart contracts
  • Created showcase platform for artists to display and promote their work
  • Collaborated in team environment to deliver winning solution at competitive hackathon
TECHNOLOGIES
ICP Motoko Blockchain NFT Smart Contracts

playashesi - Ashesi Sports Department Website

Full-Stack Developer (Group 14 - webtechbros)

2025

Comprehensive sports management system for Ashesi University's Sports Department. Built with Next.js 16, React 19, TypeScript, and Supabase PostgreSQL database. Visit project →

  • Developed fixture and score management system with sport-specific scoring (goals, sets, points, quarters) and automatic standings calculation
  • Implemented team and player management system with roster management, match participation tracking, and player statistics
  • Built role-based access control system with Row Level Security (RLS) policies for Sports Chairman, Score Masters, Team Managers, and Viewers
  • Created public information display system with fixture browsing, match details, standings, and news/announcements
  • Designed 3-tier architecture (Presentation, Application, Data) with Next.js App Router, Supabase API, and PostgreSQL database
  • Implemented real-time score updates, automatic standings calculation, and comprehensive match statistics tracking
TECHNOLOGIES
Next.js React TypeScript Supabase PostgreSQL RLS Tailwind CSS

Ledger - Multi-Chain Crypto Exchange Backend

Backend Engineer

Production-Ready

A production-ready, multi-chain cryptocurrency exchange backend built with NestJS and TypeScript. Supports non-custodial wallets, real-time price feeds, buy/sell functionality, P2P transfers, and Paystack fiat deposits.

  • Core Features: Multi-chain support (Bitcoin, Ethereum, Solana), non-custodial wallets, real-time price feeds from CoinGecko, CoinCap, and Binance, NGN fiat buy/sell trading, P2P transfers, and Paystack-powered fiat deposits
  • Security: JWT authentication with refresh token rotation, TOTP-based 2FA with QR setup, withdrawal address whitelisting, rate limiting, Helmet security headers, and strict DTO request validation
  • Technical Architecture: WebSocket layer for real-time prices and notifications, BullMQ workers for blockchain operations, Redis caching for prices/sessions, PostgreSQL with Prisma ORM, and auto-generated Swagger/OpenAPI documentation
TECHNOLOGIES
NestJS TypeScript PostgreSQL Prisma Redis BullMQ WebSockets Swagger/OpenAPI JWT 2FA Paystack

Ledger - Secure Crypto Wallet Frontend

Frontend Engineer

Multi-Chain Wallet

A comprehensive cryptocurrency wallet frontend built with React, TypeScript, and modern web tooling. Provides a secure and intuitive interface for managing multi-chain wallets, trading, transfers, and NGN fiat payments.

  • Multi-Chain & Wallets: Unified interface for Bitcoin, Ethereum (via Wagmi/RainbowKit), and Solana wallets with balances, history, and whitelisted addresses
  • Trading & Transfers: In-app swaps with real-time quotes, order history, external address transfers, Web3 wallet signing, gas estimation, and detailed transaction status tracking
  • Fiat & Security: Paystack-based NGN deposits, payment verification, bank account management, JWT auth with refresh tokens, 2FA, encrypted local storage, CSP, rate limiting hooks, and confirmation flows
  • UX & Realtime: Modern responsive UI using shadcn/ui and Tailwind, WebSocket-driven price updates, interactive portfolio charts, toasts, loading and error states, and dark mode support
TECHNOLOGIES
React TypeScript Vite React Router Redux Toolkit React Query shadcn/ui Tailwind CSS Framer Motion Recharts Wagmi RainbowKit Solana Web3.js BitcoinJS Socket.io Axios Crypto-JS Zod React Hook Form Paystack

Drop - P2P File Sharing Platform

Full-Stack Engineer

Elixir Phoenix · WebRTC

Instant, secure, peer-to-peer file and text sharing that runs entirely in the browser. Built with Elixir Phoenix and WebRTC, featuring zero installation, zero file size limits, local-network optimization, and end-to-end encryption readiness.

  • Core Experience: Browser-only PWA with no installs, P2P WebRTC transfers that never touch persistent servers, automatic Relay/Bridge fallback, hotspot-friendly QR onboarding, shared clipboard, live device discovery, and real-time chat.
  • Performance & Security: AES-256-GCM-ready E2E encryption with room passwords, SHA-256 integrity checks, 4x faster chunked transfers (16KB → 64KB), live speed/ETA stats, drag-and-drop anywhere, and offline-ready service worker.
  • Roadmap & Extensibility: Planned UX upgrades (previews, batch uploads, smart icons, verification badges) and advanced features like folder transfer, REST API, and programmatic room/device/transfer management.
TECHNOLOGIES
Elixir Phoenix Phoenix LiveView WebRTC Simple-Peer Phoenix PubSub Phoenix Presence Tailwind CSS PWA Service Workers Bandit

Tone - AI-Powered Foundation Matching

Founder & Frontend Engineer

Production-Ready · Q2 2026

An intelligent, premium web application that uses advanced AI to detect skin tone and recommend perfectly matched foundation products, with a luxurious, commerce-ready UI.

  • AI Skin Analysis: Machine learning–powered skin tone detection from camera input, tuned for diverse complexions and lighting conditions.
  • Live Camera & Try-On: Real-time face detection with MediaPipe, interactive virtual try-on using color overlays with adjustable hue and intensity.
  • Smart Product Matching: Foundation recommendations from a structured makeup database, with tone clustering and shade ranking.
  • Premium UX: Dark-mode first experience with glassmorphism, smooth transitions, authentication and onboarding flow, and fully responsive mobile-first layout.
TECHNOLOGIES
Next.js TypeScript React Tailwind CSS MediaPipe OpenCV Computer Vision AI/ML Authentication PWA

ICPUARMS - Integrated Computer Program University Academic Records Management System

Full-Stack Developer

Desktop Application

A comprehensive Windows desktop application for managing university academic records, built with Visual C++ and MySQL. Provides role-based access for students, faculty, interns, and administrators to manage courses, enrollments, grades, attendance, fees, and evaluations.

  • Student & Course Management: Complete student profiles, enrollment tracking, course creation with prerequisites, scheduling, and academic records management
  • Grade & Attendance System: Flexible grade weight configurations, grade items, automated GPA calculations, and date/session-based attendance tracking
  • Fee Management: Fee creation, payment processing, receipt generation, and financial record keeping
  • Course Evaluations: Multi-question type evaluations with anonymous response support and comprehensive analytics
  • Reporting & Transcripts: Automated transcript generation with semester and cumulative GPA, graduation tracking, and comprehensive reports in multiple formats (PDF, CSV, JSON, XML)
  • Security & Access Control: Role-based access control (RBAC) for all user types, secure authentication, password management, and data integrity enforcement through foreign keys and validation rules
  • Architecture: Three-tier architecture with Windows Forms UI (Presentation), domain models and business rules (Business Logic), and MySQL database with repository pattern (Data Access)
TECHNOLOGIES
Visual C++ C++/CLI Windows Forms MySQL MySQL Connector C++ Visual Studio Desktop Application RBAC Three-Tier Architecture
Public Health

An AI-powered platform detecting mercury and chemical poisoning through AI, community data, and environmental sensing. Addresses the public health crisis in Ghana where over 100,000 people are exposed to mercury and heavy metals through contaminated air, water, and soil. Visit project →

  • AI-Powered Detection: Advanced machine learning models detect mercury and chemical poisoning in humans, plants, and the environment by analyzing symptoms, images, and sensor data to identify early warning signs before severe damage occurs
  • AI Chat Bot: Interactive assistance with text and voice support in local dialects, providing instant help for health queries and contamination concerns
  • Evidence Upload: Picture upload or live camera capture functionality for documenting contamination evidence and environmental hazards
  • Community-Driven Reporting: Empowers citizens to report contamination incidents, building a comprehensive database of environmental risks
  • Real-Time Risk Mapping: Dynamic visualization of pollution hotspots with interactive maps and data graphs showing contamination patterns across regions
  • Water Quality Testing: Automated water quality monitoring integration to track contamination levels and provide early warnings
  • Smart Alerts & Consultations: Real-time risk alerts enable swift intervention and support, connecting affected communities with healthcare resources
TECHNOLOGIES
Machine Learning Generative AI Computer Vision Natural Language Processing Voice Recognition Real-Time Mapping Environmental Sensing Data Analytics Public Health

EPL Match Prediction App

Machine Learning Engineer

Machine Learning

A machine learning application that predicts English Premier League (EPL) match outcomes using historical data from 2014-2023. The app provides win probabilities and predicted scores for matches between any two EPL teams.

  • Match Outcome Prediction: Predicts the result (Home win, Away win, or Draw) for any EPL match using machine learning models trained on historical performance data
  • Score Prediction: Estimates likely scorelines based on historical performance patterns and team statistics
  • Win Probabilities: Calculates probability percentages for each possible match outcome (Home win, Away win, Draw) with confidence metrics
  • Interactive UI: User-friendly Streamlit interface with team logos and EPL styling for seamless user experience
  • Historical Data Analysis: Uses over 10 years of EPL match data (2014-2023) for accurate predictions and model training
  • Real-time Predictions: Instant prediction generation when teams are selected, providing immediate insights for match analysis
TECHNOLOGIES
Python Streamlit Scikit-learn Pandas NumPy Joblib PIL Machine Learning Data Analysis

Tennis Match Prediction App

Machine Learning Engineer

Machine Learning

A comprehensive machine learning application that predicts tennis match outcomes using extensive historical data spanning from 1877 to 2023. The app determines which player or duo will win in both singles and doubles matches, leveraging over 140 years of tennis match data for accurate predictions.

  • Comprehensive Historical Data: Utilizes tennis match data from 1877 to 2023, covering over 140 years of professional tennis history for robust model training
  • Singles & Doubles Prediction: Predicts match outcomes for both individual players and doubles teams, handling different match formats and team dynamics
  • Player Performance Analysis: Analyzes historical performance patterns, head-to-head records, and player statistics to inform predictions
  • Win Probability Calculation: Determines which player or duo will win with confidence metrics based on comprehensive data analysis
  • Machine Learning Models: Trained on extensive historical datasets to identify patterns and trends in tennis match outcomes across different eras
  • Temporal Analysis: Accounts for changes in tennis rules, playing styles, and player performance evolution over the decades
TECHNOLOGIES
Python Machine Learning Scikit-learn Pandas NumPy Data Analysis Historical Data Predictive Modeling

Algorithmic Trading System

Developer

July 2024 – Present

Developed a trading system utilizing Metatrader, TradingView, TD Ameritrade, that generates trading opportunities based on specific market conditions, optimising decision-making for traders.

  • Implemented algorithms to determine the optimal duration for holding positions, enhancing trade profitability
  • Provided traders with actionable information and insights to support profitable trading strategies and minimise risks resulting in cumulative positive returns
TECHNOLOGIES
Python TradingView Metatrader Data Analytics

Social Media Sentiment Analysis System

Developer

2024

Created social media sentiment analysis system for Nestlé using Natural Language Processing, LLMs, and API integration for data collection

  • Implemented API integration and data collection systems for social media platforms
  • Applied Natural Language Processing and Large Language Models for sentiment analysis
  • Delivered strategic insights improving customer engagement and product development
TECHNOLOGIES
NLP LLMs API Integration Python

Semantic Search Platform

Developer

2024 – Present

Implemented semantic search capabilities and data processing pipelines for multiple AI-driven applications

  • Developed semantic search functionality using vector embeddings and similarity matching
  • Built data processing pipelines for efficient information retrieval and ranking
  • Integrated RAG (Retrieval-Augmented Generation) capabilities for enhanced AI applications
TECHNOLOGIES
RAG Semantic Search Vector Embeddings Python

Buddy Up - Peer Mentorship Platform

Full-Stack Developer

2024

Built scalable peer mentorship platform with multi-step registration system

  • Implemented real-time tracking and gamified rewards system
  • Developed full-stack application using modern web technologies
TECHNOLOGIES
Full-Stack React Real-time

Ucart - E-commerce Platform

Full-Stack Developer

2024

Built e-commerce platform enabling African users to purchase from global stores

  • Implemented alternative payment methods and mobile money integration
  • Created seamless shopping experience for cross-border transactions
TECHNOLOGIES
E-commerce Payment Integration Mobile Money

Bus (Pegasus) - Smart Transportation System

Full-Stack Developer

2024

Developed smart transportation system with real-time tracking and booking interface

  • Implemented route optimization for campus and city commutes
  • Created user-friendly platform for efficient public transportation management
TECHNOLOGIES
Real-time Tracking Route Optimization Transportation

BookMe - Hair Business Management Platform

Full-Stack Developer

2024

Created platform for hair businesses featuring clientele management and inventory tracking

  • Implemented booking systems that increased revenue by 25%
  • Developed comprehensive business management solution for salon operations
TECHNOLOGIES
Business Management Booking System Inventory

Yango Accra Mobility Prediction Hackathon

Machine Learning Engineer - Finalist

July 2025

Finalist in Accra Taxi Prediction hackathon, developing predictive models for ride time optimization using trip and weather datasets

  • Applied advanced feature engineering techniques and machine learning algorithms, achieving strong RMSE performance on competitive leaderboard
  • Collaborated in high-pressure environment demonstrating problem-solving, data analysis, and model reproducibility under strict evaluation criteria
TECHNOLOGIES
Machine Learning Feature Engineering Predictive Modeling

Design Services

UI/UX Designer & Visual Designer

Ongoing

Creating stunning app UI, posters, and videos for various clients. Leveraging design abilities to make UI solutions user-driven and intuitive.

  • Designing on Figma and Canva for various client projects
  • Creating stunning app UI designs that prioritize user experience and usability
  • Developing posters and marketing materials for brand promotion
  • Producing video content for client campaigns and presentations
  • Leveraging design abilities to make UI solutions user-driven and intuitive
  • Ensuring all designs align with client brand identity and user needs
TECHNOLOGIES
Figma Canva UI/UX Design Visual Design Video Production Graphic Design