The person behind the brand
Cross-domain practice, made to actually ship
The founder of Luce Agentic has 5+ years of hands-on experience across finance, Web3, and consulting. Started Luce Agentic in 2026 with one goal: make AI more than a demo — make it a reliably delivering system inside real workflows.
Not coming out of an AI research lab — coming out of the actual friction of frontline work.
The starting point of my career was a management trainee program at a large financial institution, where I saw "organization-level workflow" complexity through decision cadence and approval chains. Later I moved into Web3 product and consulting, accumulating 5+ years of cross-domain experience in financial services, Web3 products, and advisory work — all facing the same problem: tools keep multiplying, but workflows never actually get simpler.
AI changed the resolution angle. But the gap between "usable" and "shippable" is still wide — most teams stall in the trial phase. Starting Luce Agentic in 2026 isn't because AI is hot; it's because I know how to make it actually run inside workflows.
Based in Taipei, serving clients remotely across the US, EU, and Singapore — closing the gap between "bought the AI tool" and "shipped the AI workflow."
- 10+
- clients served
- 10+
- projects delivered
- 3
- continents (US · EU · SG)
Clients range from independent algorithmic traders, to e-commerce and trade automation, to AI one-person companies built from zero — delivered across everything from personal productivity to cross-team collaboration.
- Completed all official Anthropic courses
- Claude Meetup Lightning Talk speaker
- AI-tools instructor for government & enterprise
- National Chengchi University — Information Management
- Financial-industry professional certifications
- Luce Agentic columnist on Medium
Implementation before theory.
I don't start with a whiteboard framework. I start by asking what your workflow actually looks like today. For every AI embedding point, the question is "can this ship?" — not "does this look smart?"
Three principles support the design of every Luce Agentic service:
- 01
Smallest viable first
Find the smallest version that runs, then iterate outward.
- 02
Workflow before tool
Tools are means. Workflow design is the core.
- 03
Delivery as the milestone
Every engagement has a measurable "shipped" endpoint.
I don't believe there's a one-size-fits-all formula for AI adoption. The AI workflows of finance contexts and the design logic of Web3 products are completely different. The complexity of personal automation and cross-team collaboration aren't even on the same order of magnitude.
So what Luce Agentic does is: understand your workflow first, then design the AI embedding strategy that best matches your reality — instead of dropping a ready-made template on top and calling it "done."