Backend Dev · ML Engineer · Designer
Available for work

I'm Isaac Adejuwon

Backend developer from Nigeria. Building scalable products, exploring ML in public, and designing experiences that connect.

IA
Projects
View All →
MV
MartVille
E-commerce storefront · Backend API
SP
Stress Prediction ML App
Physiological signals · Apple Watch
DP
Diabetes Prediction Model
SVM · XGBoost · Random Forest
WP
LeadLight & Client Sites
WordPress · PHP · Custom backend
PIS
Protradeec Intergrated Services
PHP · Custom backend
Skills
Backend
PythonPHPREST APIsDatabasesGit
Machine Learning
Scikit-learnXGBoostSVMData Analysis
Design
Adobe SuiteFigmaBranding
At a Glance
3+
Years building
10+
Projects shipped
2
ML models
Learning in public
About

I'm Oluwasegun Isaac Adejuwon — a backend developer, ML engineer, and creative designer from Nigeria.

I build scalable backend systems, train ML models on real-world problems, and bring it all together with thoughtful design.

Part of the Creatives' Cohort community — sharing my journey openly.

Computer Science degree

Foundation in software engineering

Backend development

Python, PHP, APIs, databases

ML engineering

Stress & diabetes prediction models

Creative design

Adobe Suite, Figma, branding

Contact

Let's build something great.

Open to freelance, collaborations, and full-time backend or ML roles.

Type E-CommerceBackend
Stack PythonREST APIPostgreSQL
Status ✅ Shipped
Year 2024
MV
MartVille

A persistent e-commerce storefront built to replace WhatsApp status selling — giving sellers an always-on product listing, order management, and customer connection without technical overhead.

Visit Project →
The Problem

WhatsApp statuses disappear after 24 hours. Sellers had to re-upload products daily, wasting time and losing buyers in the process.

The Solution
  • Persistent storefront live 24/7
  • Product listings with images & prices
  • Scalable API for multiple sellers
  • No tech knowledge required
Tech Stack
PythonREST APIPostgreSQLBackend Architecture
Outcome

Sellers gained a permanent storefront replacing WhatsApp selling. Product discovery improved with always-on listings and a structured browsing experience.

← No previous
Next →
Stress Prediction
Type Machine LearningHealth Tech
Stack PythonScikit-learnApple Watch
Status ✅ Model trained
Year 2024
SP
Stress Prediction ML App

Predicts stress levels using physiological signals — heart rate, EDA, and skin temperature — with an Apple Watch integration concept for real-time wearable monitoring.

View Model →
The Problem

Real-time stress detection tools are expensive or clinical. There was a gap for an accessible, data-driven monitor built on everyday wearables.

The Solution
  • ML model on physiological signals
  • Heart rate, EDA, temperature inputs
  • Apple Watch integration concept
  • Multi-class stress classification
Tech Stack
PythonScikit-learnPandasNumPyApple Watch API
Outcome

Trained classification model for physiological stress detection, bridging ML research with wearable health tech via Apple Watch architecture.

Type Machine LearningHealth AI
Stack SVMXGBoostRandom Forest
Status ✅ Deployed
Year 2024
DP
Diabetes Prediction Model

A comparative ML study evaluating SVM, XGBoost, Random Forest, and Logistic Regression for early diabetes detection — benchmarked on accuracy, precision, recall, and F1.

View Demo Streamlit → View Demo Render →
The Problem

Early diabetes detection saves lives but requires the right algorithm. A rigorous multi-model comparison was needed before deployment.

The Approach
  • Cleaned clinical diabetes dataset
  • Trained 4 classification models
  • Compared all key metrics
  • Best model selected for production
Tech Stack
PythonScikit-learnXGBoostSVMRandom ForestMatplotlib
Outcome

XGBoost achieved top accuracy and was prepared for deployment as an accessible early-detection tool.

Type Web DevClient Work
Stack WordPressPHPCustom Backend
Status ✅ Live
Year 2023–2024
LL
LeadLight & Client Sites

Custom WordPress builds with backend and frontend integration for real clients — translating each brand into a polished, performant, and maintainable web presence.

Visit Site →
The Problem

Clients needed professional sites that reflected their brand without custom-code complexity — fast, maintainable, and on-brand.

The Approach
  • Brand discovery & identity mapping
  • Custom WordPress theme in PHP
  • Backend: forms, CRM, API integrations
  • Performance optimization & SEO
Tech Stack
WordPressPHPCustom ThemesJavaScriptCSSSEO
Outcome

Live, client-approved websites with custom integrations — delivered on time and within scope.