Software Engineer & ML Explorer — Akure, Nigeria

Oyebamiji
Mustapha.

Building systems that scale across continents.

I write backend software and train neural networks. Currently studying Computer Engineering at FUTA and shipping features that reach schools across 10 countries. I care about clean architecture, real performance numbers, and code that other humans can read.

Read my CV
5Countries
3Hackathon wins
4+ Backend Frameworks
3 Industry Roles
Mustapha Oyebamiji

A bit about
who I am.

I'm Oyebamiji Mustapha Oyetunde — people call me Musoye. I'm a fourth-year Computer Engineering student at Federal University of Technology, Akure, and I've been writing software seriously since 2022.

My work lives mostly in the backend — distributed systems, microservices, cloud infrastructure, and the plumbing that keeps large platforms running. Lately I've been building a federated learning framework for medical imaging, which has been one of the more interesting problems I've worked on.

Outside code, I lead bootcamps at the Google Developer Student Club on campus, tutored thermodynamics and discrete maths for peers, and try to write the occasional technical article when I have something worth saying.


What I work with

Languages, frameworks, databases, and tools I reach for regularly.

Languages

PythonJavaScriptTypeScriptGoPHPC / C++SQLBash

Frameworks

DjangoFlaskExpress.jsLaravelVue.js

Databases & Cloud

PostgreSQLMySQLMongoDBDynamoDBOpenSearchRedisAWS (S3, SQS)Google Cloud

ML / AI

PyTorchTensorFlowscikit-learnOpenCVPandas / NumPyTransformersFederated Learning

DevOps & Tools

DockerLinuxNginxHAProxyGitREST APIsGraphQLMicroservices

Experience

Where I've worked and what I built there.

Sep 2025
— Present

SchoolTry · Switzerland

FullStack Engineer

  • Architected platform-wide audit logging using Laravel — capturing sensitive admin actions across 60+ schools in 10 countries.
  • Led migration from legacy permission system to team-scoped RBAC with zero downtime across 60+ institutions.
  • Built full-stack features in Laravel + Vue.js covering result management, payment reconciliation, and institutional admin.
  • Improved data integrity by implementing granular access controls, reducing risk of unauthorized record manipulation.
LaravelVue.jsPostgreSQLRBAC
Jan 2025
— Present

Nexora · Contract

Backend Engineering Lead

  • Led backend architecture for multiple client platforms using Node.js, designing services for high-concurrency workloads.
  • Engineered secure payment pipelines via Paystack API and a blockchain asset verification platform using Moralis.
  • Managed versioned REST API deployments on AWS ensuring clean microservice synchronization.
Node.jsExpress.jsPaystackAWS
Jan 2025
— Apr 2025

Peeriva

Backend Engineer

  • Built a real-time Notification Service with Redis + Amazon SQS — 10K+ messages/day at 99.9% delivery reliability.
  • Developed a User Activity Service with DynamoDB, cutting latency 55% across 100K+ activity logs.
  • Orchestrated a distributed Search Service integrating OpenSearch, PostgreSQL, and DynamoDB — 3× faster queries.
DjangoRedisAmazon SQSOpenSearch

Honors &
Achievements.

A record of things that went well.

25

Intelligent Planet Hackathon

KFUPM · Saudi Arabia · Top 25 of 2000+ global teams

Built EcoFlow AI — an event-driven digital twin on Google Cloud processing real-time IoT data to predict crowd surges and optimise carbon emissions. Selected from over 2,000 global applications.

2nd

Wema Bank Hackathon

Fintech · Security Track

Built TrustLock — a secure gateway for remote companies to monitor and authenticate employees, with digital auth, user priority levels, and IP/location/device tracking.

8th

Squadco Hackathon

EdTech / Fintech · 8th of 108 teams

Developed SSFH — a platform for automating school fee savings for students, including backend services to handle transactions efficiently at scale.

Hackathon moment
Award moment
Hackathon moment

Research

Academic work currently in review.

In Review · 2025

Lightweight and Privacy-Preserving Federated Deep-Learning Framework for Efficient Brain Tumor Detection Across Distributed Medical Centers

Mustapha O., Charles O.

A federated learning framework for collaborative MRI analysis without sharing patient data. Achieved 84% accuracy with a custom CNN and 98.5% with EfficientNetB0, optimised via pruning and quantisation so it runs on standard clinical hardware — not just research clusters.


Leadership & Teaching

Roles beyond writing code.

Bootcamp & Skills Development Lead

Google Developer Student Club (GDSC), FUTA — 2023/24 & 2024/25

Organised and led coding bootcamps introducing students to backend development, cloud technologies, and real-world project experience. Mentored participants through hands-on technical work.

Tutor — Engineering Thermodynamics (MEE 206)

Muslim Student Society of Nigeria, FUTA

Tutored up to 30 students, helping them improve understanding and exam performance.

Peer Tutor — Discrete Maths & Algorithms / C Programming

CPE 201 & CPE 202, FUTA

Mentored departmental peers in core computer science subjects. High pass rates among mentees.

Projects &
Work.

Things I've built, shipped, or researched.

Backend · FinTech · Voice

Padipay

Voice-controlled payment system supporting English and Nigerian English. Multilingual voice recognition, real-time transaction processing, and secure backend validation.

Python · Django · MongoDB · SpeechRecognition · PostgreSQL

2nd · Wema Bank Hackathon

Fullstack · Security

TrustLock

Secure workforce access gateway — RBAC, JWT auth, and full audit trails tracking IP, device identity, and geolocation. Built for distributed remote teams.

Django · PostgreSQL · Redis · JWT · Vue.js

Top 25 / 2000+ · Saudi Arabia

ML · IoT · Cloud

EcoFlow AI

Event-driven digital twin on Google Cloud processing real-time IoT data to predict crowd surges and optimise carbon emissions. Firestore for real-time, BigQuery for analytics.

Flask · MySQL · Google Cloud · Firestore · BigQuery · Nginx

Publication Pending

ML · Medical AI · Research

Brain Tumor Detection

Custom CNN + fine-tuned ResNet for MRI tumour classification. Extended into a federated learning framework achieving 98.5% accuracy across distributed hospitals without sharing patient data.

PyTorch · TensorFlow · OpenCV · Flask · Federated Learning

ML · AgriTech · Mobile

CropWatch

AI crop disease detection via transfer learning on CNNs. Two-stage training — feature freezing then fine-tuning. Mobile backend for real-time field inference from leaf images.

Flask · MySQL · Nginx · CNN Transfer Learning

ML · NLP · LLM

Fine-tuning GPT-2 for Sentiment

Rebuilt GPT-2 in PyTorch and fine-tuned on a custom dataset for high-accuracy sentiment classification. Tokeniser setup, model optimisation, and reduced inference latency.

Python · PyTorch · Transformers · HuggingFace

Backend · AI · ALX Final Project

HelpBot

AI document assistant that parses PDFs and answers contextual queries via OpenAI. Response time reduced by 65% through prompt tuning and payload optimisation.

Node.js · Express.js · OpenAI API · Vue/Vite

ML · FinTech

Fraud Detection System

ML model flagging fraudulent transactions using monetary data patterns, with a Flask web interface for simulation and real-time prediction workflows.

Python · scikit-learn · Flask · Pandas

ML · Computer Vision

Facial Emotion Recognition

Detects human emotions from facial images — happy, sad, angry, surprised. Web app for uploading photos and getting real-time predictions.

Python · OpenCV · TensorFlow · Streamlit