Inside the R&D Lab:
Applied AI & Deep Tech

Where theory meets execution.

Currently based in Taiwan, I work with local inference and agentic tools as well as cloud AI infrastructure like Google Vertex AI. I have built out Linux-based NVIDIA multi-GPU workstations and work with the Apple MLX layer in addition to traditional software design and development for practical applications. This hands-on work keeps my advisory grounded in reality, not whitepapers.

Active Projects & Sandboxes

Project 01

The Automation Backend

Modular LLM Engine. A self-improving automation layer where I test multi-agent orchestration, vector databases, and dynamic code generation workflows in real-time.

Tech Stack & Focus

  • • Multi-agent orchestration frameworks
  • • Vector databases & RAG pipelines
  • • Dynamic code generation
  • • Local inference optimization

Project 02

Cognitive Interfaces & UX

Custom LMS Platforms. Proprietary spaces where I experiment with human engagement optimization, cognitive retention, and UX architectures across domains like industrial training, compliance, and education.

Tech Stack & Focus

  • • Human-in-the-loop UX design
  • • Cognitive retention algorithms
  • • Real-time adaptive learning paths
  • • High-performance frontend frameworks

Project 03

The Consumer Edge

Aggregate Web Apps. High-traffic consumer tools built specifically to stress-test real-time API latency, model drift, and cost-efficiency at scale across global networks.

Tech Stack & Focus

  • • Edge computing & CDN routing
  • • API latency optimization
  • • Model drift monitoring
  • • Cost-efficiency at scale

Project 04

Biomedical Computation

Deep Tech Moonshot. Building the data pipelines and AI models required to analyze, halt, and reverse spinal bone fusion and autoimmune inflammation (specifically Ankylosing Spondylitis).

Tech Stack & Focus

  • • Medical imaging data pipelines
  • • Autoimmune inflammation modeling
  • • High-throughput compute (Multi-GPU)
  • • Apple MLX & NVIDIA CUDA

Project 05

Autonomous Agent Orchestration

Enterprise Sandbox. A dedicated environment for testing complex, multi-step agentic workflows that interact with legacy enterprise systems without causing disruption.

Tech Stack & Focus

  • • Legacy system API bridging
  • • Agentic workflow safety rails
  • • State management & rollback
  • • Google Vertex AI integration

Want to see these systems in action?

I leverage this internal R&D lab to build out functional proof-of-concepts for my advisory clients. If you need to stress-test a concept before committing heavy engineering resources, let's talk.

View Advisory Offerings