Theophylus Maboshe

I'm a full-stack developer and Computer Science student who loves building complete systems from scratch, not just the parts that are fun, but the whole thing: native mobile apps, web dashboards, backends, and the AI/automation layer that ties it all together.

What I'm good at

• Native mobile: Kotlin/Jetpack Compose on Android, Swift/SwiftUI on iOS. I've shipped apps with media playback, image caching, and push notifications.
• Web: React, Vite, and Tailwind for fast, responsive dashboards with rich text editors, charts, and resumable file uploads.
• Backend & data: Supabase, PostgreSQL, and hand-written Row Level Security policies so data privacy is enforced at the database level, not just the app layer. When I need more than a BaaS can give me, I build custom Node.js/Express services, or reach for Rust (tokio, Axum) when performance really matters.
• AI & automation: Local LLM deployment (Qwen, Gemma, and others), AI agents, and Python/Shell scripts that automate the repetitive stuff so I can focus on building.

What drives me

I like taking an idea from a blank folder to something people actually use, and I think about projects as complete systems rather than a pile of scripts. I get excited by unfamiliar tech; if I don't know a tool yet, I'll dig in, learn it, and fold it into whatever I'm building. I'm especially drawn to AI that does things, agents that plan, execute, and recover, rather than AI that just chats.

Interesting things I've built

• A full church management ecosystem: native Android and iOS apps, a React admin dashboard, a Node.js API, and a Supabase/PostgreSQL backend with RLS, all designed and shipped as one connected product.
• Rogue Quant: an AI-driven trading bridge connecting MetaTrader 4 to a locally-hosted 27B-parameter LLM, with a Rust/Axum backend, a custom SSE parser to stream the model's reasoning live, and a hacker-themed WebSocket dashboard.
• Cinero: a Python-based gold trading analysis system with multi-timeframe technical indicators (RSI, EMA, MACD) and backtested Buy/Sell/Hold signal generation.
• Nova: an autonomous AI computer agent that plans, executes, verifies, and asks for approval before taking significant actions in a terminal environment.
• A Solana wallet monitoring system in Go, and a WhatsApp Business automation project exploring chatbot workflows and SaaS-style deployment.

I'm always looking for the next thing to build end to end.