Business Intelligence Engineer · Data Systems Developer · AI Product Builder

I build automated reporting systems, AI-powered products, predictive analytics models,
and production-grade data pipelines that turn raw data into decision intelligence.

I specialize in building scalable analytics systems and AI-powered products that automate insight generation, support data-driven decision-making, and solve real problems for real users.

✔ AI-powered products deployed to production
✔ End-to-end BI systems (SQL → Python → KPI Engine → Dashboard)
✔ Machine learning models for forecasting & risk prediction
✔ Automated reporting pipelines (PDF + Email + Scheduling)
✔ WhatsApp bots & conversational AI systems
✔ Decision-support systems for institutions

Live product. Real users. Built with ₦0.

A fully deployed AI-powered WhatsApp chatbot built for Nigerian market traders and small business owners. Understands natural Pidgin English and handles complete business operations through WhatsApp — no app download required.

What it does:

• Records single and multiple sales in one natural message
• Tracks stock in real time with automatic deduction on sales
• Manages customer debts and payment history
• Generates daily, weekly, monthly, and custom date range summaries
• Analyses top products by revenue, profit, and ROI
• Tracks top customers by purchase volume
• Supports multiple business sections (food, drinks, clothes) separately
• Runs 24/7 on a live server without human intervention

Tech: Python · Flask · Groq AI (LLaMA 3.3) · Twilio WhatsApp API · Supabase · Render
🔗 View Repository · 🌐 Live Server

Production-grade BI pipeline. Zero manual effort.

End-to-end automated analytics system that extracts 2,800+ real sales transactions, computes 12 executive KPIs, and delivers formatted PDF reports to stakeholders via scheduled email — every day without human intervention.

What it does:

• Extracts live transactional data from SQL Server
• Computes 12+ executive KPIs across 19 countries and 7 product lines
• Generates structured multi-page PDF reports automatically
• Sends HTML email briefings with attachments on a daily schedule
• Deploys a 5-tab interactive Streamlit dashboard with YoY comparisons, heatmaps, and choropleth maps

Tech: Python · Pandas · SQL Server · Plotly · Streamlit · ReportLab · smtplib
🔗 View Repository · 🌐 Live Dashboard

95% model accuracy. 37,000+ records.

Trained ML models on historical institutional data to forecast student enrolment and optimize resource allocation for academic planning.

What it does:

• Trained supervised learning models on 37,000+ historical records
• Achieved 95% prediction accuracy
• Built an interactive prediction dashboard for institutional planners
• Supports proactive resource allocation and planning decisions

Tech: Python · Scikit-learn · Pandas · Streamlit

Predicts at-risk students before dropout occurs and identifies key risk drivers using supervised learning — enabling proactive intervention by academic institutions.

Churn modeling and revenue impact quantification for telecom customer data. Identifies high-value customers at risk and quantifies potential revenue loss.

• Built and deployed a live AI product serving real Nigerian traders with ₦0 infrastructure cost
• Reduced manual reporting time by 90% through full pipeline automation
• Enabled real-time KPI tracking for executive decision-makers
• Achieved 95% accuracy in enrolment forecasting for institutional planning
• Designed scalable architectures that separate data ingestion, transformation, and presentation layers

I build intelligent data systems and AI-powered products that solve real problems at scale. My goal is to work at the intersection of data engineering, machine learning, and product development — creating systems that don't just analyze data but act on it.

• AI Product Development roles
• BI Engineering & Data Engineering opportunities
• Data Analyst positions
• Freelance & International collaboration

📧 thesamueloyedokun@gmail.com  |  🔗 Portfolio  |  🔗 LinkedIn

"The analysis is the starting point. The system is the product. The impact is the proof."