🧭Project

Here's what I actually built, ran, or shipped — and what happened when I did.

August 2025 – October 2025

Most people wait until they are already in the country to start job hunting. I ran the entire search from Taipei — no UK network, no UK education, no engineering background, and no right to work until my visa cleared.

The goal was to run it like a product: define the variables, isolate them one by one, test fast, and iterate.

Phase 1: Lay the groundwork

Before sending a single application, I eliminated every variable I could control.

Phase 2: Build the networking engine

Target: at least 1 coffee chat per week, max 3.

Phase 3: Validate the CV and test job titles

Phase 4: Fix the interview

CV was not the issue. Getting interviews was not the issue. Failing them was.

Phase 5: Expand the funnel

Available startup job listings were limited — not enough volume to hit two and a half offers. Opened a second channel.

Phase 6: Win the assignment rounds

First offer in month one. Two and a half offers in month two (yes, half — writing an article about it) — while still based in Taiwan, with no UK education, no UK work history, and no engineering background. Accepted the Founder Associate role at Seal.

Skills: Cold outreach · LinkedIn prospecting · Funnel building · Interview scripting · A/B testing · Networking systems · Market research · Visa navigation

2025 – Present

I used to think building products required either learning to code or hiring someone. AI changed that assumption completely. Starting in 2025, I began using AI as a hands-on collaborator, not just for writing or research, but for actually shipping things.

What surprised me was not that AI could help. It was how fast the feedback loop became. I could go from idea to deployed product in days instead of months.

1. This Portfolio Site: https://portfolio.hankwhy.com/

I built a fully custom portfolio site using notablog, a Node.js static site generator that pulls content directly from Notion as a headless CMS, deployed on Netlify. AI handled the debugging, configuration, and edge cases I would have been stuck on for weeks.

2. Hugo Blog Migration: https://hankwhy.com/

I had two years of writing scattered across Medium — on a platform I did not control. I migrated everything to a self-hosted blog built on Hugo, deployed on Netlify.

Two sites shipped. Full control over my digital presence. The bigger outcome was the habit: I now default to building instead of waiting. AI closed the gap between having an idea and having something real.

Skills: Prompt engineering · AI-assisted development · Static site generators (notablog, Hugo) · Netlify deployment · Content architecture

January 2025 – Present

I started an anonymous Threads account in January 2025 to practice writing in public without the pressure of personal branding. The constraint was intentional — if the writing did not stand on its own, nothing else would carry it.

I was not consistent. No schedule, no growth strategy, no optimization. I posted when I had something worth saying about tech, VC, startups, and whatever I was reading.

1,500+ followers without a single piece of promotion. The account is still running. It validated something I suspected: the ideas matter more than the name behind them.

Skills: Content curation · Writing in public · Idea distillation · Audience building

GoFreight · 2025

GoFreight sells logistics SaaS to freight forwarders — a global, fragmented, relationship-driven market. The sales team knew their ICP in theory but spent hours each week figuring out who to actually call. There was no systematic answer to the question: who are our best next 400 customers?

I built the system to answer that question at scale.

Cut prospecting time from days to hours. Every rep started each week with a clear, prioritized list. It became the foundation the sales team ran on — and the first time outbound felt systematic rather than reactive.

Skills: ICP definition · Data enrichment · Market mapping · Scoring model design · Sales operations · Cross-functional alignment

GoFreight · 2024

GoFreight had a product worth selling and a growing list of freight forwarder prospects. What it did not have was a reliable, scalable way to reach them. Outbound was essentially manual — individual reps writing one-off emails with no consistency, no data, and no infrastructure underneath.

I built the system from the ground up.

40%+ open rates and 8% reply rates at peak. Industry average is roughly 20% open and 2-3% reply. Cold email became the primary top-of-funnel channel for the sales team — and the first time outbound felt like a repeatable system rather than a gamble.

Skills: Email deliverability · Domain configuration (SPF/DKIM/DMARC) · Outbound copywriting · Clay · Sequence design · A/B testing · Sales enablement

National Taiwan University · 2019

In my first year of university, I noticed that campus-branded hoodies were something students wanted but no one had made easy to buy. I decided to build a small business around it.

This was not a class project. It was a real commercial operation with real constraints — no funding, no team, limited time, and a market that had not proven itself yet.

The first challenge was demand aggregation. Individual purchases would not move enough volume to make the unit economics work, so I designed a group-buying competition across clubs and departments. The competitive element created social momentum — nobody wanted their club to lose.

From there, the GTM was scrappy by design:

Every decision was made with limited resources and a bias toward speed. The goal was to validate demand fast, move inventory faster, and reinvest into the next cycle.

NT$1M revenue in 5 months. My first experience building something real — spotting demand, designing GTM tactics, and executing without a safety net. It is also where I first understood that scrappy and systematic are not opposites.

Skills: Product design · Pre-order validation · Supplier negotiation · Grassroots marketing · Inventory management