I am a full-stack engineer and R&D specialist with a strong focus on real-time audio processing, LLM-based automation, and immersive 3D/VR experiences.
My work spans low-level system programming, cross-platform audio capture, shader development for VRChat, and building AI-driven pipelines that enhance both productivity and creativity.
I explore the boundaries between real-time systems, generative AI, and interactive environments, developing practical tools such as ProcessAudioTap (per-process audio capture across Windows/macOS/Linux), VRChat automation utilities, and advanced AudioLink-driven visual effects.
My engineering philosophy emphasizes performance, clarity, and reliability—designing systems that run fast, scale smoothly, and integrate elegantly with complex pipelines. Whether building audio drivers, multimodal AI pipelines, GPU shaders, or end-to-end backend systems, I deliver solutions that are technically deep and user-centric.
- Python, TypeScript/JavaScript, C#, C/C++, Rust (basic)
- React / Next.js, Node.js, FastAPI, Electron
- Unity (C#), ShaderLab, HLSL
- WASAPI / CoreAudio / PulseAudio / PipeWire
- Real-time PCM manipulation (EQ, filters, mixing, VAD)
- Low-latency audio pipelines, streaming, WebRTC integration
- Whisper (OpenAI, Faster-Whisper), pyannote.audio
- RAG pipeline engineering
- LLM orchestration (OpenAI, Claude, Gemini)
- Custom TTS models (ElevenLabs, XTTS)
- AWS (EC2, S3, CloudFront, Lambda, Redshift Serverless)
- GCP (Cloud Run, Vertex AI, Cloud Functions)
- Docker, GitHub Actions, PyPI trusted publishing
- CI/CD, monitoring, observability
- VRChat world design & automation
- AudioLink-driven shaders
- Unity URP built-in rendering pipelines
- Real-time post-processing & distortion FX
Cross-platform per-process audio capture library (Win/macOS/Linux).
Features low-latency audio taps, real-time filters, CLI utilities, plugin architecture, and VRChat integration.
Custom shaders reacting to audio spectra, used in club events, DJ worlds, and visual installations.
Unified workflow that integrates Whisper → LLM → Slack/Discord bots.
Supports diarization, summarization, topic detection, and automated posting.
Tools for avatar detection, join-log analysis, OSC automation, and AudioLink-reactive effects.
Marketing/CRM analytics platform with AI-powered predictive modules, built with Next.js, Python, and Redshift.
Architected backend, admin tools, data pipelines, and AI integrations.
- GitHub: https://github.com/m96-chan
- X (Twitter): TBD
- Booth: TBD
- VRChat: TBD
- LinkedIn: TBD