Machine
Learning
Supervised and unsupervised models with scikit-learn, TensorFlow and PyTorch — from feature engineering to evaluation.
Hello Everyone!
I'm
ML Engineer · Python Specialist · Full Stack Developer
I build intelligent systems that ship. From data pipelines and model training to the API and interface that put a model in front of real users — designed for accuracy, scale and maintainability.
Python
TensorFlow
FastAPI
Machine
Python
Engineering
Clean, tested, well-typed Python. OOP design, Pandas and NumPy pipelines, and performance profiling that holds up in production.
Full Stack
Development
FastAPI, Django and Flask on the back end, React and Vue on the front — one coherent product, not stitched-together parts.
Data
Engineering
PostgreSQL and MongoDB schema design, indexing and query optimisation, plus ETL pipelines that keep models fed with clean data.
APIs &
Microservices
Async REST services with validation, versioning and sane error handling, ready to integrate with third-party platforms.
Security &
Deployment
JWT auth, role-based access control, containerised deploys and CI so what works locally works everywhere else.
About Me
Turning models into products people use
I'm Eunice Muturi, a machine learning engineer and Python specialist based in Kenya. Most ML work stalls somewhere between a promising notebook and a system anyone can actually rely on — closing that gap is the part I enjoy most.
I take a problem from data exploration and model selection all the way through the API, the database and the interface, then leave behind documentation and tests so the next person isn't guessing. I care about correctness first, clarity second, and clever last.
End-to-end ownership
Production-minded ML
Clear communication
0Projects shipped
0STI model test accuracy
0Tools & frameworks
0Remote-ready delivery
My Latest Projects
Eight builds from my GitHub — predictive models, multi-tenant backends, e-commerce platforms and the interfaces on top of them. Every card links to a case study and the real source.
.png)
Machine Learning
STI Predictive Model
Clinical risk-scoring platform: a Random Forest classifier served through a Django Ninja API, with a React dashboard, county heatmap and full audit trail.
.png)
Machine Learning
Loan Default Risk Predictor
KNN and Naive Bayes trained side by side to classify Kenyan loan applicants into three risk bands, with cross-validated tuning and a Django interface.
.png)
Backend
WhatsApp Commerce Platform
Multi-tenant SaaS backend: 15 Django apps, JWT auth, Celery and Redis, M-Pesa Daraja and the WhatsApp Business API, with encrypted per-tenant credentials.
.png)
Full Stack
FindAHome
Real estate platform with four roles, admin verification gates and a fraud moderation queue. Django REST behind a React 19 and TypeScript front end.
.png)
Full Stack
Kilimo Guru
Smart agriculture platform for Kenyan farmers: crop records, weather alerts, a produce marketplace, M-Pesa payments and an offline-capable PWA.
E-commerce
Craft Cove
Django marketplace for handmade goods: categories, reviews, wishlists, a cart that survives login, and reserved stock released when a payment fails.
E-commerce
TJ Naturals
Skincare store with nested categories, SKU-level inventory and low-stock thresholds, zone-based shipping rates and a support chatbot, on Docker Compose.
Full Stack
Vehicle Management
Car hire platform with role-separated accounts, date-range bookings, generated booking references and automatic duration-based pricing.
No projects in this category yet.
Also on GitHub: AI Evaluation Portfolio — structured technical reviews comparing AI-generated Python solutions against test suites and design criteria.
Technical Proficiency
Where my day-to-day hours actually go.
Machine Learning & Data Science
Full Stack Development
Core Technologies
FAQs
Mostly supervised learning on structured data — classification, regression and forecasting where the payoff is a clear business decision. I handle the full pipeline: cleaning and feature engineering, model selection and tuning, honest evaluation, then deployment behind an API so the predictions actually reach users.
Yes — that's the reason I work full stack. I build the FastAPI or Django service, design the PostgreSQL schema behind it, and put a React or Vue interface on top. You get one system from one person, instead of coordinating hand-offs between three.
A focused model or API usually lands in two to three weeks. A full application with an ML component runs closer to six to eight. I scope it properly after the first call rather than guessing, and I share progress in weekly increments you can actually run.
Remote by default and comfortable joining an existing codebase, repo conventions and review process. I write code meant to be read by other people — typed, tested and documented — so onboarding your team onto it is part of the delivery, not an afterthought.
You keep the source, the documentation and a recorded walkthrough. Models drift, so I offer a retainer for monitoring, retraining and new features — but it's optional, and nothing is built in a way that locks you into me.
Testimonials
Feedback from people I've built with.
PL
Product Lead
Startup Collaboration"Eunice delivered clean backend architecture quickly and communicated technical decisions clearly throughout the project."
TM
Team Mentor
Developer Program"Great ownership and consistency. The final product was responsive, polished, and ready for production use."
The stack I build with
every day
Let's talk
Python
TensorFlow
FastAPI
React
PostgreSQL
Docker
Git
Currently Exploring
Deepening my work in advanced ML techniques, production ML systems and scalable architecture with modern DevOps practice.
PyTorch
Kubernetes
Advanced FastAPI Patterns
MLOps
Contact Me
Open to new projects, collaborations and full-time opportunities.