Nairobi, Kenya

Ruth Mutile Kimeu

Data scientist and machine learning practitioner who came to data through a decade inside institutions that could not afford to break. I build the predictive models, deployed AI tools, and reporting systems that turn scattered operational data into decisions organisations can defend.

Data Scientist ML Practitioner Product Builder Systems Designer Founder, BriefDesk Solutions
284,807
Transactions modelled
in fraud detection pipeline
0.860
Average precision
FraudLens, tuned threshold
98%
ALX Africa ML track
final examination
10+
Years institutional
operations and analytics
About

Where data science meets operational reality

My work sits at the intersection of data science, operations, and institutional systems. The hard part is rarely the model. It is the messy data and the decision around it.

For over a decade I worked inside organisations that could not fail: hospitals preparing for JCIA, ISO, and CAP accreditation, and a university network coordinating across East Africa, Central Asia, and Europe. That environment taught me that the bottleneck is almost never intelligence or effort. It is scattered data, manual reporting, and systems no one trusts.

I now solve that as a data scientist. At Aga Khan University's Network of Quality Teaching and Learning, I manage reporting systems, quality frameworks, and institutional documentation, and I led a migration to automated Power BI dashboards that cut reporting cycle time by 40 percent.

Through BriefDesk Solutions, my consultancy, I help schools, NGOs, and institutional teams move from manual workflows to clean data, working dashboards, and decisions they can act on, including deployed AI products now in production.

I am completing a BSc in Data Science at the Open University of Kenya, and I finished the ALX Africa Machine Learning track at 98 percent. I also hold a BSc in Business Studies and an Associate Degree in Health Sciences.

InstitutionsQuality frameworks, reporting, and audit-ready documentation
SchoolsOperations systems, dashboards, and continuity planning
SMEs & NGOsFinance tools, M&E reporting, and digital products
Finance & RiskFraud, credit, and decision models grounded in business cost
01 / Modelling

Python & Machine Learning

pandas, NumPy, scikit-learn, XGBoost. Classification, regression, NLP, recommender systems, model evaluation, and imbalance-aware methods.

02 / Data

SQL & Data Engineering

Relational schema design, multi-table joins, CTEs, data cleaning pipelines, and structured analysis in MySQL and PostgreSQL.

03 / Decisions

Dashboards & BI

Power BI, DAX, Google Sheets systems, KPI and executive reporting that leadership can actually read and act on.

04 / Products

AI Applications

LLM-powered tools and chatbots, prompt design, and deployment with Next.js, Streamlit, and the Claude API.

05 / Systems

Systems Architecture

Multi-file linked operations systems, automation, and workflow design that gives organisations data visibility.

06 / Governance

Quality & Writing

Institutional review frameworks, accreditation support, gap tracking, and technical reports, proposals, and case studies.

BriefDesk Solutions

Structure. Systems. Smarter decisions.

The consultancy side of my work: helping organisations move from scattered workflows and manual reporting to clearer systems, data products, and better decisions.

AI Solutions

Deployed AI tools

Chatbots, report automation, and document-aware assistants, with a live reference in production.

Analytics & BI

Dashboards & reporting

Data cleaning, KPI reporting, Power BI, and decision support for leadership.

Operations

Systems & OMS builds

Linked Google Sheets systems, workflow design, and systems audits.

Governance

SOP & quality

SOP packs, governance frameworks, gap tracking, and accreditation support.

Portfolio

Selected projects

Models, AI products, analyses, and operational systems built for real finance, institutional, education, and SME contexts.

ML · Classification
Rare-outcome modelling
Probability calibration
2025–2026

Draw Hunter 3000 — Calibrated Rare-Outcome Prediction

A draw-specialised Random Forest classifier predicting football draw outcomes across 1,752 matches in five European leagues. A study in modelling a difficult, low-base-rate class with engineered features and calibrated probabilities. Selected for the ALX Tambali Portfolio Spotlight.

77%
Draw recall
0.683
AUC-ROC
1,752
Matches
Pythonscikit-learnRandom ForestFeature Engineeringpandas
Data Science · Agriculture
Predictive modelling
2025

Maji Ndogo Agricultural Yield Analysis

An end-to-end analysis of 1,000 farm records across seven provinces and nine crop types, identifying soil moisture, fertiliser use, and irrigation as primary yield drivers and building a regression model to estimate yield on held-out records.

1,000
Farm records
0.966
R² (held-out)
1.216
MAE
Pythonpandasscikit-learnRegressionEDA

Structured educational dataset; results reported on held-out validation.

Product · Fintech
AI finance tool
Launching soon

BiasharaFlow — SME Finance Tracker

A WhatsApp-first, AI-powered finance tracking tool for Kenyan SMEs, simplifying daily income and expense tracking through mobile-first workflows and natural language input. Registered with the Kenya Copyright Board.

AI IntegrationFintechProduct DesignSME Tools
Systems · Education
Operations design
Deployed

TERA School Operations Management System

A 12-file linked Google Sheets system managing attendance, fees, academics, staff, and inventory for a primary school in Vihiga County, Kenya, with a supporting institutional website.

Google SheetsIMPORTRANGEDashboard DesignData Architecture
Writing & Thought Leadership

Published author and practitioner

Books on administrative leadership and institutional practice, registered with the Kenya Copyright Board, plus an ongoing newsletter on data-driven operations.

Book · 2026

The Invisible Architecture

Administrative Leadership as Institutional Power. How governance systems, data intelligence, and operational design drive organisational excellence.

ISBN 978-9914-27-341-0
Book

The Person Who Keeps It Moving

On the often unseen administrative practice that holds organisations together.

ISBN 978-9914-27-340-3
Book · Forthcoming

Life After Retirement

A practical guide to navigating the transition beyond formal employment.

In preparation
How I work

From problem to decision

My work combines technical analysis with operational context, so the output is useful to both decision-makers and the people who implement it.

01

Define the problem

Clarify the operational question, decision need, users, constraints, and success criteria.

02

Structure the data

Clean, organise, validate, and transform messy information into usable datasets and workflows.

03

Analyse and model

Use EDA, statistics, dashboards, and machine learning to find patterns and test assumptions.

04

Turn insight into action

Translate findings into recommendations, dashboards, systems, and clear next steps.

Get in touch

Let's build something useful

Open to machine learning and data science roles, analytics and contract work, systems consulting, and institutional data strategy, across East Africa and internationally.

Typical response: 24–48 hours