I founded Bowdler and shipped a native macOS AI video-processing app that performs on-device speech recognition with zero cloud calls. It reached 82 trial downloads and 40 paid users through a $59 one-time purchase.
I built its multi-engine ASR pipeline around Whisper MLX, Parakeet GGUF, Silero VAD, and Vosk, reaching 3.63% WER on an internal benchmark. I also engineered Secure Enclave hardware-bound licensing, a Go/PostgreSQL licensing API, and a Swift AVFoundation media engine.
Before Bowdler, I worked independently on Python async systems, including a Telegram automation platform serving 60K+ users. I also delivered NFT market scraping processing 45K+ daily listings, prediction-market monitoring, and TON blockchain analytics APIs.
I work across Python, Go, Rust, Swift, and JavaScript/TypeScript, with experience in production infrastructure, native desktop apps, audio processing, and AI-assisted development.
