Project directory Case studies you can actually evaluate.
Each project below is real — shipped, in active development, in collaboration, or in paper-trade research. Status badges make the difference explicit.
Live on the App Store iOS Fitness Platform
Villain Arc Plan sessions, track sets in real time, and get actionable progression suggestions powered by deterministic training rules plus on-device intelligence.
State-driven workout lifecycle: pending → active → summary → done. AI-assisted progression suggestions with outcome tracking across sessions. SwiftUI SwiftData + CloudKit Foundation Models (on-device AI) App Intents + App Shortcuts ActivityKit Live Activities WidgetKit
Open source · self-hosted AI Transcription Platform
Transcribing App Self-hosted Whisper transcription with a one-link UI and a JSON API. Local-first by default.
From link to transcript in minutes, with live progress visible the whole way. Local-first architecture — your audio never leaves your machine. FastAPI (ASGI) Faster-Whisper (CTranslate2 runtime) yt-dlp Media Acquisition FFmpeg Audio Extraction Python 3.10+ Backend Static HTML/CSS/JS Frontend
Paper trading · evidence-first Prediction Market Research
FCT Weather Bot A weather-driven trading agent in paper-trade evaluation. Live capital deferred until the audit trail proves the signal.
Quote-aware decision loop with explicit skip conditions and sizing rules. Full paper-trade audit log: probability, EV, fill, exit trigger, settlement, result. Python Prediction Market APIs Weather Forecast Ingestion Quote-aware Decision Loop Paper Trading Audit Log
Active collaboration · paper trading Collaboration · Prediction Markets
KNYC Signal Platform A two-studio collaboration on prediction-market signal infrastructure. Paper-trading first, live capital later.
With ReSync Technology LLC (New York)
Joint build with ReSync Technology LLC, a New York-based studio, on infrastructure and strategy. Dockerized stack with a paper-trader preflight gate before any live trade. Python Docker Prediction Market APIs Shared Signal Infrastructure Make-driven Paper-trader Preflight
Open source · pipeline available AI Voice Cloning Infrastructure
Voice Pipeline Turn rights-cleared source audio into a finetuned voice model that runs locally, without cloud dependencies.
Custom voice models trained from your own audio — owned outright, no per-use cloud fees. Local-first architecture means audio and model never leave your infrastructure. F5-TTS Voice Cloning faster-whisper Transcription yt-dlp Media Acquisition FFmpeg Audio Normalization Python 3.11 Pipeline Runtime PyTorch Local Inference
