Data Scientist & Software Engineer based in Nairobi, Kenya. I build end-to-end machine learning pipelines, full-stack digital products, geo-spatial systems, and multi-tenant SaaS architectures for real-world markets.
I am in my final year of a BSc in Data Science and Analytics at Jomo Kenyatta University of Agriculture and Technology (JKUAT).
I bridge the gap between complex data pipelines and responsive, user-first software products. Beyond academia, I run Apollos Digital Solutions, a venture studio where I engineer custom tech stacks for startups, legal professionals, campaign managers, and international businesses.
Whether it's deploying fine-tuned ML classifiers, parsing highly unstructured mobile money transaction statements via automated regex patterns, or mapping demographic data visually for county campaigns, I focus on building production-grade, highly validated software.
My studio acts as a full-cycle development incubator. I've designed and shipped production systems for legal services, real-time messaging, logistics mapping, automotive OCR tracking, and US financial wealth report pipelines.
Building end-to-end ML pipelines including regression analysis, hyperparameter tuning (GridSearchCV), dimensionality reduction (PCA, L1 Regularization), and cross-validation models.
Designing pipelines that convert highly messy text and unstructured files (e.g. M-Pesa SMS strings, IEBC CSV registrations) into validated, query-ready database tables.
Creating interactive choropleth mappings and demographic analyses using GeoPandas, Folium, and Leaflet to map regional metrics and demographic trends.
Developing multi-tenant web and mobile architectures using Flutter, Next.js, FastAPI, Node.js, and Supabase with strict Row-Level Security and offline sync strategies.
Integrating LLMs and automation frameworks into active software workflows (e.g. OCR capturing, custom PDF generation via ReportLab, auto-calculators).
Developing election campaign modeling algorithms, aspirant SWOT analytics platforms, and constituency-level analytics engines using raw census and voting datasets.
Click a tag below to highlight the projects that use that technology!
Intelligent document analysis and risk engine that automates balance sheet ingestion (Gemini 1.5 Flash), runs double-entry validations, and applies localized SSA benchmarks via RAG to predict credit default risk.
Built a five-component spatial data framework for battery swap network expansion, optimal placement, and PAYG default risk prediction.
Automated quarterly financial report portal for a US wealth management firm, replacing a full-day manual canvas/word process with FastAPIs and ReportLab.
An analytics engine processing IEBC voter trends, youth ratios, and swing seats across 47 counties with interactive Streamlit maps.
Digital service marketplace connecting customers with local provider portfolios. Integrates Socket.IO, Daraja API subscription payments, and cron management.
Production-grade logistics PWA for fleet management. Multi-tenant Row-Level Security, Supabase Realtime updates, offline-first mutations using Dexie.js.
Political intelligence tool mapping 114 Nyakach polling stations. Tri-layer scenario outcome simulators, aspirant SWOT grids, and spatial voter registration targets.
Aggregated 446 polling stations dataset for gubernatorial strategy. Spatial mapping and competitive candidate metrics deployed in a custom web dashboard.
SaaS repair shop management workflow. Gate Pass protection, OCR number plate scanning, and automated WhatsApp quote formatting.
Financial tracking system. Parses raw, messy unstructured M-Pesa statements via Regex patterns into database entries coupled with Power BI tracking.
ML pipeline classifying samples. Managed curse of dimensionality (p>>n) using L1 regularization, SMOTE resampling, and GridSearchCV.
Mapped Kenya Master Health Facility List distributions and ownership dependencies across 47 counties using interactive Folium layers.
UCI Seeds analysis using K-Means, Hierarchical Clustering, and multiple classifiers compared via McNemar's hypothesis testing.
Microfinance loan risk modeller. Achieved 79% overall accuracy and 95% default precision mapping utilizing logistic regression classification.
NCPB government agricultural data mining for the 2025 Long Rains season, plotting planting to top-dressing ratios via Plotly dashboards.
Investigated funding allocations and startup trajectories utilizing K-Means grouping and PCA feature reduction tools.
Practice management engine for advocates compliant with Kenya Data Protection Act 2019. Matter billing, billing schedules, and AI brief summaries.
Offline-first capstone mobile platform designed for remote community health workers, featuring zero-latency structured database queries.
Lead Software Engineer (Remote)
Developed a dual-system luxury travel ecosystem. Integrated a public-facing website and a private concierge portal. Managed via a monorepo setup, built with Vercel edge endpoints and continuous GitHub Actions deployments.
AI Automation Consultant (Atlanta, USA)
Designed and built a FastAPI financial portal automating quarterly investment dashboards. Generates pixel-perfect ReportLab PDFs for Simple Automated Cashflow Systems (SACS) and Total Client Charts (TCC), saving the firm full days of manual graphic editing.
Lead Data Strategist
Synthesized county voter registration CSVs spanning 446 polling stations. Engineered geographic models identifying voter splits, delivering ward maps, slides, and spreadsheet pricing models for electoral executives.
Digital Infrastructure Engineer (Nairobi)
Modernized digital practices for a Nairobi-based law firm. Architected their core web portal, secure document handling pipelines, and legal research workflows.
Data & Systems Lead
Structuring voter analytics dashboards mapping regional support. Managing real-time target profiles and candidate competitiveness monitors for tactical allocation.
Jomo Kenyatta University of Agriculture and Technology (JKUAT)
Relevant coursework: Machine Learning I & II, Time Series, Artificial Intelligence, Stochastic Processes, Regression Modelling, Tests of Hypothesis, Theory of Estimation, and Software Engineering.
Power Learn Project (PLP)
Immersive full-stack engineering program covering core web structures (Python, SQL, JavaScript, React) and cross-platform mobile frameworks (Dart, Flutter).
ALX Africa
Professional training focused on virtual assistance, task tracking, agile workflows, and stakeholder communication.
Kisumu Boys High School
Graduated with a Mean Grade of B+.
I am seeking full-time roles, junior engineering opportunities, freelance engagements, or attachment contracts commencing June 8, 2026.