Data Scientist building production ML systems and full-stack platforms, with hands-on experience spanning fraud detection modeling, LLM/RAG engineering, and end-to-end web infrastructure for a multi-country sports organization.

I'm a BSc Data Science graduand at JKUAT (Jomo Kenyatta University of Agriculture and Technology, Karen Campus), currently serving as Tech Lead of ASK IN SPORTS, a youth football and basketball tournament organization operating across Kenya, Uganda, Rwanda, and South Sudan. I work at the intersection of data science and software engineering — building models that need to hold up under real-world scrutiny, and platforms that need to stay up under real-world traffic.

• 🎓 BSc Data Science, JKUAT
• 💼 Tech Lead, ASK IN SPORTS (askinsports.org)
• 🌱 Currently completing the DataTalks.Club LLM Zoomcamp, building a RAG-based Q&A assistant as capstone
• 📫 Email: macoutmatur94@gmail.com
• ⚡ I like problems that sit between clean math and messy production systems

FMCSA-compliant trip planning app for commercial truck drivers, built with Django REST Framework and React (TypeScript).

• Engineered an Hours-of-Service simulation engine enforcing 11-hr drive limits, 14-hr duty windows, and 70-hr/8-day cycle caps, with automatic break and rest insertion
• Generates interactive route maps and DOT-format ELD log sheets
• Integrated OpenStreetMap Nominatim for geocoding and Leaflet.js for interactive mapping
• Deployed via GitHub CI/CD — frontend on Vercel, backend on Railway

Full CRISP-DM pipeline built on the Kaggle Credit Card Transactions Fraud Detection Dataset.

• Engineered features including Haversine distance and temporal transaction patterns
• Trained and compared four models — Random Forest (PR-AUC 0.8798), XGBoost (0.8795), Logistic Regression, and LightGBM
• Deployed as a Streamlit app with masked card display, tier-based transaction limits, three-model consensus voting, SHAP explainability, and SQLite persistence
• Delivered with a full JKUAT-formatted academic report

End-to-end web platform for a multi-country youth sports organization, built and maintained as Tech Lead.

• Migrated infrastructure to Firebase with Cloudflare deployment (Vite + @cloudflare/vite-plugin)
• Built a multi-step team registration flow, Cloudinary image pipeline, and EmailJS-powered contact system
• Developed an interactive Leaflet impact map and Firestore security rules
• Built a full admin panel: Firebase Auth, Recharts dashboards, CRUD management, CSV export
• Currently extending to a React Native (Expo) mobile admin app

End-to-end ELT pipeline built on a synthetic banking dataset (9 CSVs, ~5.8M rows total), fully containerized in Docker.

• Modeled data through a medallion architecture — raw → staging (typed, cleaned, constrained) → marts (star schema)
• Built a star schema with 6 dimension tables and 4 fact tables spanning transactions, cards, loans, and support tickets
• Ran on a 3-container Docker Compose stack (Postgres, pgAdmin, Jupyter) with idempotent, re-runnable ETL notebooks
• Handled real-world type-casting edge cases (numeric-looking text fields, integer-encoded booleans, legitimate nulls)
• Delivered analytical insights on transaction volume, fraud rate by card type, and loan pricing patterns

Coursework-driven ML engineering across the DataTalks.Club LLM Zoomcamp.

• Built a vector search pipeline using ONNX embeddings
• Ran a retrieval evaluation pipeline (Hit Rate, MRR) across 295 chunks
• Built a full RAG-based course assistant with OpenTelemetry tracing and Grafana/Postgres observability, served via Streamlit
• Currently building a RAG assistant over ASK IN SPORTS content as capstone, with a DuckDB ingestion pipeline

Intern — 4G Capital

• Analyzed business data to support decision-making through BI reporting
• Built and evaluated machine learning models for predictive tasks
• Developed applications using Google ADK
• Supported management of structured business data systems

Tech Lead — ASK IN SPORTS

• Own technical strategy and delivery across web, mobile, and data infrastructure for a youth sports organization operating in four East African countries
• Lead full-stack development, cloud deployment, and admin tooling

Languages & Data

Web & Cloud

ML & Tools

Open to work involving:

• Machine learning / predictive modeling
• LLM and RAG systems
• Full-stack web platforms (React, Firebase, Cloudflare)
• Data engineering pipelines

📧 Email: macoutmatur94@gmail.com