Featured Research
TB Classification on Cough Audio
Designed a novel deep learning system to detect Tuberculosis (TB) from cough audio. Pioneered a competitive dual-autoencoder framework to classify samples. Engineered a complete data pipeline in Python to process raw audio, extract Mel-Spectrograms, feed spectrograms into autoencoders and then a classification layer.
- PyTorch
- Python
- Autoencoders
- Spectrograms
- Google Colab
Experience
Full Stack Web Developer · AfriPaw
Developed a smart web app for a non-profit using Next.js, TypeScript, and Prisma. Implemented user authentication, automated notifications, and scheduled database backups to Google Cloud Storage.
- Next.js
- TypeScript
- PostgreSQL
Data Engineering Intern · SpatialEdge
Built profitable Forex trading algorithms in Python. Developed an anomaly detection model for trading data using an Autoencoder in Keras and researched financial forecasting technologies.
- Python
- Keras
- Anomaly Detection
Other Projects
Crypto Arbitrage Flashloan Bot
Developed a crypto arbitrage bot on Ethereum using Flash Loans from Balancer. Authored Solidity smart contracts and a TypeScript bot to scan for and execute profitable price discrepancies.
- Solidity
- Hardhat
- Ethers.js
Cryptology Educational App
Developed a React Native application to accompany a university cryptology textbook. Implemented encryption/decryption ciphers (RSA, ElGamal) and cryptanalysis techniques.
- React Native
- TypeScript
- Algorithms
What To Cook Recipe App
Developed a dynamic recipe discovery platform using the T3 Stack (Next.js, TypeScript, tRPC, Prisma). Implemented a personalised recommendation engine using a weighted scoring algorithm based on user preferences and available ingredients.
- T3 Stack
- Prisma
- NextAuth
- Tailwind CSS
Technical Arsenal
Languages
TypeScript, Python, C++, Java, Solidity, SQL, HTML/CSS
Frameworks
Next.js, React, TensorFlow, PyTorch, Tailwind CSS, Hardhat
Tools
Git, PostgreSQL, Firebase, GCP, REST APIs
Concepts
Machine Learning, Smart Contracts, Anomaly Detection, Full-Stack
