ISP Financial Model - Unit Economics in Code
An interactive Python/Jupyter financial model for a Nigerian ISP business: ~15 tunable parameters (NGN pricing, data caps, COGS, capex, location capacity) driving live KPI and break-even visualizations.
An interactive Python/Jupyter financial model for a Nigerian ISP business: ~15 tunable parameters (NGN pricing, data caps, COGS, capex, location capacity) driving live KPI and break-even visualizations.
A pandas/scikit-learn CLI that turns raw Kuda neobank statements into spending analytics: category and counterparty breakdowns, 3-sigma anomaly detection, and a 30-day cash-flow forecast - built for the messy reality of Nigerian bank exports.
Two 2024 prototypes probing the edges of automation: a Puppeteer-driven tool for discovering town-level communities that match business criteria, and a local GPU voice pipeline (Whisper large-v3 → XTTS v2 voice cloning) for conversational presentations.
A media-ops toolbox built to professional packaging standards: rms-scan, a PyPI-published FFmpeg loudness-QC CLI with CI and release automation; an RMS-aware spoken-watermark embedder; Whisper/Vosk shorts pipelines; waveform and PDF stamping utilities.
A tiny Typer/Rich Python CLI that auto-documents the shape of every CSV in a folder with perfectly uniform sibling READMEs - dialect sniffing, deterministic diffable output, proper src-layout packaging and console-script entry points, at 6.5 KB of code.
An AI meeting-notes pipeline that watches Dropbox folders and writes summaries via OpenAI - proper OAuth refresh flows, ffmpeg chunking around API file limits, multi-tenant config, Docker packaging - while the venture’s grander backends were honestly left as named placeholders.
2019–2024: the posting operation of a solo music brand, fully automated - content curation bots, tweet-screenshot cross-posters, crash-tolerant launchd schedulers with missed-run recovery and idempotency files - the same problem solved four times across two languages as the craft matured.
19 repositories of self-built automation (2021–2024): machine bootstrap, dotfiles portability, contact-data pipelines, HTTP test libraries, Keynote remote control - in Python, Ruby, Shell, JXA, and Objective-C.