4 systems
Production ownership
I build systems used by operations, warehouse teams, drivers, and executive meetings.
Data Science / Production Analytics / Applied AI
I build AI and analytics systems that survive real operations.
My work spans production dashboards, OCR automation, executive reporting, mathematical modelling, and competition-grade machine learning across African business, language, climate, and finance problems.
My career evidence portfolio
Glenn Kipanga
I am a data scientist and business intelligence analyst with a mathematics foundation, production system ownership, and ranked applied machine learning work.
46
My evidence files
4
Production systems
30+
Zindi competitions
9
Top-10 finishes
8 domains
Applied AI breadth
My competition and project work spans NLP, forecasting, finance, climate, agriculture, OCR, and speech.
5 education records
Mathematical depth
My BSc Mathematics and Modelling Processes background gives me depth in ODEs, stability analysis, optimization, and simulation.
30+ competitions
Public GitHub signal
I keep my public GitHub data connected so visitors can inspect my public projects and repository activity.
Production Systems
Real tools I built for live operational work.
I see these as some of my strongest examples: systems that moved reporting, stock control, OCR extraction, and performance management out of manual workflows.
Production SystemsFireside Group
Fuel Extractor System — OCR Pipeline & WhatsApp Integration
End-to-end fuel and maintenance management system. Drivers submit fuel and maintenance receipts; the system uses PaddleOCR to automatically extract structured data (fuel amount, cost, vendor, date) from receipt images and stores it in Supabase.
Impact
Eliminated manual receipt data entry. Created an auditable, searchable fuel and maintenance dataset. Enabled real-time odometer tracking across the vehicle fleet.
PaddleOCRPythonSupabaseWhatsApp APIEvolution APIOCRfine-tuning
Reporting AutomationFireside Group
Weekly Executive Reporting & PowerPoint Automation
Every week, produced automated executive PowerPoint presentations for the Fixed Data Services and Rollout departments. Each deck contained approximately 23 charts summarizing departmental performance and KPIs.
Impact
Eliminated manual chart creation and slide building. Ensured consistent, reproducible weekly reporting across two departments with zero manual formatting effort.
PythonPowerPoint automationGoogle Sheets APIData VisualizationMatplotlibPandaspython-pptx
Production SystemsFireside Group
Operations Scoreboard — Performance Analytics Dashboard
Real-time operations performance dashboard tracking technicians, dispatchers, and Project Managers (PMs) across daily, weekly, monthly, and quarterly timeframes.
- Daily view: used in morning standup meetings for live operational decisions
- Weekly/monthly/quarterly views: trend analysis and performance management
- Per-role breakdowns: technician, dispatcher, and PM metrics tracked separately
- Clean data layer: structured schema ensures consistent, reliable performance figures
Impact
Gave leadership real-time visibility into operational performance. Replaced manual performance tracking with a live, queryable system used at executive morning meetings.
Supabasedatabase designweb developmentdashboard developmentPythonSQLdata modeling
Production SystemsFireside Group
Warehouse Management System — Full Stack Web Application
Full-stack warehouse management web application managing the complete lifecycle of materials — from arrival at the company through storage to issuance to staff.
- Material intake logging with quantity and supplier tracking
- Real-time stock level monitoring
- Issuance administration — controls and records who gets what materials
- Full audit trail of material flow from receipt to use
Impact
Replaced manual stock tracking. Created full material lineage from delivery to consumption. Gave operations management real-time visibility into stock levels and material usage patterns.
PythonSupabaseweb developmentfull stackdatabase designinventory managementREST API
Career Timeline
My through-line: mathematics, analytics, then production systems.
My path moves from academic modelling depth into healthcare analytics, consulting-style ML, and live systems at Fireside.
Fireside Group
Data Analyst — BI & Analytics Systems
Collaborate with product, finance, and operations teams to design and maintain analytics-ready datasets supporting business intelligence systems and decision-making.
Power BISQLPythonETLData ModelingPostgreSQLgitbash
Boston Consulting Group (BCG X) via Forage
Digital Transformation Data Science Virtual Experience
Completed a series of structured data science tasks simulating real BCG X client engagements. Each task provided a dataset and a business problem — the objective was to select the most appropriate algorithm, build a model, evaluate it rigorously, and translate findings into ex...
PythonSQLRandom ForestPower BIPandasScikit-learnmodel evaluationfeature engineering
Dedan Kimathi University of Technology
Fourth Year Coursework — BSc Mathematics and Modelling Processes
Machine Learning (CCS 4102) Functional Analysis (SMA 4126) Partial Differential Equations II (SMA 4145) Mathematical Biology (SMA 4174) Advanced Optimization (SMA 4236) Fourier Series and Transforms (SMA 4131) Data Analytics II (SMA 4143) Environmental Systems Modelling (SMA 4...
Machine LearningData AnalyticsSimulationProbabilistic ModellingAdvanced OptimizationEnvironmental ModellingMathematical BiologyFourier Analysis
National Hospital Insurance Fund (NHIF), Kenya
Data Science Intern
Analyzed healthcare claims datasets to support fraud detection, data quality assessment, and operational reporting at Kenya's national health insurer.
PythonSQLData ValidationDashboard DevelopmentAnomaly Detection
Applied AI Lab
My competition-grade experimentation across practical domains.
I use my Zindi archive as a credibility layer: ranked results in NLP, forecasting, climate, finance, agriculture, OCR, speech recognition, and responsible AI.
My domain coverage
I work broadly, with applied AI at the center.
Analytics14
NLP8
Climate7
Finance7
Agriculture6
Computer Vision2
Healthcare1
Telecommunication1
Top 0.6%
To Vaccinate or Not to Vaccinate
Rank 2 of 323 - NLP
Top 0.7%
COVID-19 Tweet Classification Challenge
Rank 2 of 279 - NLP
Top 1.6%
Ghana's Indigenous Intel Challenge — Rainfall Forecasting
Rank 9 of 564 - Agriculture
Top 4.5%
Measuring What Matters Proposal Challenge by ITU
Rank 4 of 88 - Climate
Top 5.5%
Loan Default Prediction Challenge
Rank 28 of 513 - Finance
Top 7.4%
Rhea Soil Nutrient Prediction Challenge
Rank 17 of 230 - Agriculture
Top 8.7%
Adaptation Atlas Data Storytelling — Track 5: Rainfall
Rank 2 of 23 - Climate
Top 14.5%
Barbados Lands and Surveys Plot Automation Challenge
Rank 26 of 179 - Computer Vision
Mathematical Foundation
My modelling degree sits behind the analytics work.
My education gives the analytics work real mathematical weight: calculus, linear algebra, stochastic processes, numerical analysis, optimization, machine learning, simulation, and independent research.
Final-year research project
Final Year Research Project — Mathematical Epidemiology of Soft Drug Use
A deterministic compartmental mathematical model examining the spread of soft drug use (alcohol, tobacco, marijuana) in a population. Divided the population into three compartments: Susceptible (S), Drug Users not in treatment (U1), and Drug Users undergoing treatment (U2).
- Formulated a system of three coupled ordinary differential equations (ODEs) governing population transitions between compartments
- Derived the Basic Reproduction Number (R0) using the next-generation matrix approach — the threshold parameter determining whether drug use becomes endemic
- Proved positiveness and boundedness of solutions to ensure biological feasibility
- Conducted local stability analysis using the Jacobian matrix evaluated at drug-free and endemic equilibria
- Proved global asymptotic stability of both equilibria using Lyapunov functions
- Performed sensitivity analysis to identify which parameters most significantly influence R0 and therefore epidemic outcomes
Year 1
2022Quantitative reasoning, mathematical proof, programming fundamentals, probability theory, statistical inference basics, linear systems and matrix operations, calculus for c...
Year 2
2023Advanced probability and statistical inference, algorithm design and complexity analysis, ordinary differential equations for dynamic systems, mathematical modelling method...
Year 3
2024Stochastic modelling for random systems, numerical methods for solving complex equations, optimization theory and linear programming, data analytics pipeline fundamentals,...
Year 4
2025Formal machine learning theory, probabilistic simulation for complex systems, advanced optimization techniques, environmental and biological system modelling, Fourier metho...
Evidence-Based Skills
Skills I can back up with evidence.
Across my work, education, projects, and competitions, the strongest repeated signals are Python, applied ML, BI, SQL, feature engineering, NLP, and data systems.
How I position my work
Production analytics
Dashboards, Supabase-backed systems, audit trails, operational decision tools.
Applied machine learning
Classification, regression, forecasting, OCR, ASR, NLP, feature engineering.
Business intelligence
Power BI, executive reporting, KPI pipelines, stakeholder-ready narratives.
Mathematical modelling
ODEs, stability analysis, stochastic processes, optimization, simulation.
41
Python
21
Pandas
19
Scikit-learn
11
feature engineering
7
classification
7
EDA
5
data visualization
5
HuggingFace
5
NLP
5
Power BI
5
SQL
5
storytelling
4
PyTorch
4
regression
4
time series
3
computer vision
3
database design
3
Supabase
Public GitHub
My public repositories, now in the right context.
I still keep GitHub updated daily, but I no longer depend on public repos alone to explain my career. They now support the broader evidence portfolio.
Synced Sep 5, 2026GitHub API live data
Applied Data SystemAug 2026
Darth Glennium
Data Scientist & BI Analyst · Mathematics foundation · Production system owner
PythonStreamlitPower BIPandasscikit-learnNLP
Applied Data SystemJul 2026
ebola seird intervention optimizer
This project implements a complete SEIRD (Susceptible-Exposed-Infectious-Recovered-Dead) compartmental model with explicit hospitalization and safe burial mechanisms. It combines:
PythonTypeScriptPowerShellShellHTMLReact
Applied Data SystemJun 2026
Darth Glennium.github.io
A public project by Glenn Kipanga.
HTML
Applied Data SystemJun 2026
sierra leone spatial demographics
An end-to-end reproducible data engineering and geospatial analysis pipeline implemented in R to clean, validate, and analyze subnational demographic trends across the 14 official administrative level 2 (Admin 2) units of Sierra Leo...
CSV
Credit Risk MLMay 2026
Mpesa Loan Default Model
This project is a machine learning-based credit risk scoring system that predicts the probability of loan default using alternative financial data extracted from M-Pesa mobile money statements.
PythonStreamlitPandasNumPyscikit-learnMatplotlib
NLP AnalyticsMar 2025
Tala Sentiment Analysis Dashboard
This project analyzes user reviews for the Tala app to understand customer sentiment. Using Natural Language Processing (NLP), the system classifies reviews as positive, neutral, or negative and visualizes the insights in an interac...
PythonExcelPandasNumPyNLTKscikit-learn
Market IntelligenceMar 2025
competitor price monitoring dashboard
This Power BI Dashboard provides insights into loan interest rates, approval times, and default risks across different lenders in Kenya. It helps financial analysts and decision-makers track competitor pricing trends and assess lend...
Power BIDAXCSV
Civic Data AppFeb 2025
Crime Sentencing
A web app, built using Dash and Plotly, that allows users to explore crime sentencing and statistics in Kenya.
PythonDashPlotly
Conversational AIFeb 2025
Prototype 2
This is a conversational chatbot built using Flask and deployed on Render. It can interact with users and provide responses based on a set of predefined rules.
PythonFlaskNLPRender
Contact
Let's build data products that hold up under scrutiny.
I'm open to AI, analytics, automation, dashboard, and data systems work where clean thinking matters as much as clean code.