A platform by Shuli Li, CPA. Exploring the intersection of capital strategy, systems architecture, and the frontier of AI-driven hardware — through the lens of someone who has built forecast frameworks from the ground up, and lived in the operational complexity of both US and China financial systems.
Real problems, real frameworks. Each project applies financial and systems thinking to questions that matter — from tax optimization to AI hardware economics.
Built this entire platform — design, copy, deployment — through a single Claude conversation, without writing a single line of code. The process itself was the experiment: learning that the bottleneck in AI-native building isn't technical, it's clarity of vision. Every decision that made this site distinctive came from asking the right questions: What is this not? Who is it for? What story does only I get to tell?
From blank canvas to shulili.com live: one session. Tools used: Claude, GitHub, Vercel, Namecheap.
Not sure how much to convert to Roth this year? Enter your income, capital gains, and IRA balance — the model automatically finds the lowest-tax conversion window. No appointment. No $400/hr consultation fee. Built by a CPA.
Traditional forecasting treats error as noise to be minimized. Control systems treat error as signal to be fed back. This project maps the feedback loop logic of servo control onto rolling forecast methodology — what is the financial equivalent of a PID controller?
When Tesla's BOM fell, the market repriced everything. What does a finance professional see that an investor misses? A deep-dive into how Bill of Materials curves, vertical integration decisions, and capital allocation efficiency interact.
Finance doesn't just measure outcomes — it shapes how decisions are made. These pieces explore AI hardware and frontier technologies through a financial lens: capital allocation, cost structures, and the constraints that define what can scale.
The humanoid robot isn't just an engineering bet — it's a capital allocation decision. What matters isn't whether it works in isolation, but whether it scales under a viable cost structure. Read through ROIC, depreciation curves, and vertical integration, Optimus becomes less about robotics — and more about whether Tesla can compress cost faster than complexity grows.
A working hypothesis: AI hardware startups don't struggle only with R&D — but with forecasting under uncertainty. When BOM is unstable and production is evolving, small errors compound into capital misallocation. From a finance perspective, the challenge isn't precision — but building models that adapt faster than reality changes.
Focused on how financial constraints shape technological outcomes.
The rarest professionals are not specialists or generalists — they are integrators. A mind shaped by high-altitude mountains, cross-cultural systems, and a relentless curiosity for how things actually work.
Cloud Rest, July 5th — my first time going that far, that high. I didn't know how hard it would be. I walked until my body was done. But I kept going.
Then came Half Dome.
I used to think fear was a stop signal — the cables, the exposure, the 900-foot drop on both sides. But somewhere on that climb, something shifted. Fear didn't disappear. It lost authority.
I climbed anyway. And that became a pattern: Fear doesn't get the final word — action does.
Every time something looks impossible from the bottom, the real challenge isn't the climb. It's how you interpret the fear.
"原来跨越恐惧的瞬间，就是自信诞生的时刻。"
"The moment you cross fear is the moment confidence is born."
5 of my top 10 themes fall in the Strategic Thinking domain — a rare configuration for a finance professional. Ideation, Futuristic, and Strategic show up every time a problem needs a framework that doesn't exist yet.
Operating at the intersection of US and China financial systems isn't just technical — it requires understanding how culture shapes assumptions about reporting, risk, and what "transparency" actually means.
Seven years ago, I was settling securities trades in a highly regulated banking environment — where precision wasn't optional, and process defined everything. That's where I learned something that stayed with me: precision and structure don't limit creativity — they enable it.
Since then, I've worked across accounting and FP&A — building forecasting models, owning $30M+ capital programs, and navigating cross-border finance in clinical biotech environments. But the pattern I kept noticing was this: the numbers weren't just reporting outcomes. They were revealing how the system actually worked.
That led me to a question I keep coming back to: What does the financial model reveal about how a system behaves under real constraints?
I'm drawn to environments where that question matters — AI hardware, robotics, infrastructure at scale. Places where finance isn't just tracking the business. It's helping define what's possible.
Let's Connect → LinkedIn →