Founding engineer at Replicant Trader, founder of Black Eternum Capital, and a research assistant at USIU–Africa. Writing from Nairobi.
Three current roles, one continuous workshop.
Our mission is to democratize access to high-quality trading simulations and financial tools, enabling users to build trading confidence, understand market dynamics, and enhance their strategies.
Our mission is to empower traders and investors by providing access to essential resources and knowledge tailored to Kenya’s dynamic market.
Machine Learning
I collaborate on cutting-edge research and build solutions that make a difference. Most days, the work sits somewhere between the headings below.
Two publications on one side, eight years of jobs on the other. Click any role to read what it actually involved.
Four shelves: the languages I write in, the tools on the bench, the soft work, and the stuff that's just life.
Three schools, three different rhythms.
Five undergraduate projects I'm still happy to show. Each card lists the stack and the highlights worth opening.
Individual project: built a desktop Parking Lot Management System with end-to-end ticketing, check-in/out, receipt generation, and MySQL-backed persistence. I owned system design, database integration, and core UI flows.
Login → Identification → Check-in → Level/Space allocation → Ticket/Receipt → Checkout with fee computation.
Established persistent storage, integrated MySQL via JDBC, and handled CRUD operations for tickets and sessions.
Explored ML workflows on real datasets: preprocessing, model training, and evaluation; experimented with ensembles.
Imputed missing values via NaN+median and standardized features before model training.
Compared multiple learners within ensemble frameworks; analyzed performance vs. single models.
Built Android apps covering activities, fragments, persistence (SQLite/SharedPreferences), sensors, threading, location, and UI patterns; plus VS/UWP samples. All projects were completed individually.
Hands-on implementations: forms, menus, data persistence, sensor data, threading, WebView, and location.
Collected sensor data and persisted app state/data using SQLite and SharedPreferences.
Covered OS concepts and hands-on practice: coded in C and Bash, studied CPU scheduling (Round Robin) and deadlock avoidance (Banker's algorithm), plus practical scripting labs and a led team exercise.
Implemented or evaluated Round Robin scheduling and studied/implemented Banker's algorithm concepts for deadlock avoidance.
Led a 4-person team: planned tasks, coordinated implementation, reviewed scripts, and presented outcomes.
Implemented core data structures and classic search algorithms with attention to time/space complexity and edge cases.
Built Linear, Binary, Jump, Interpolation, and Exponential search variants; documented trade-offs and optimal conditions.
Insertion and deletion at given positions; boundary checks and shifts.
Repositories, demos, half-finished ideas I keep coming back to. Every card lists the stack, the source, and where to see it running.
A selection from the Unsplash room. Twenty-something titles, four thousand uses, one stubborn habit. Tap any tile — the caption settles below.
Things on the shelf. Not the work itself, but the reminders it counted.
Twenty-something stamps across software, finance, and process.
Closed-door conferences on one side, evenings volunteered on the other.