Survived a scam interview mid-hackathon. Just submitted my project anyway.
This week I've been heads-down building ReguLink Asia for the DevNetwork AI+ML Hackathon 2026 — an AI-powered legal compliance platform for Asia-Pacific digital trade regulations.
Midway through, I got a "remote job interview" invite. Seemed legit until they asked me to run an unfamiliar ZIP package locally. Hard no. Closed it, went back to building.
Today I finally got everything across the line — application form, demo video (compressed it three times to get under the file size limit 😅), YouTube backup link, technical memo. Submitted.
What I built:
Evidence-based query engine with citation IDs and verified source URLs
Regulatory diff engine that tracks law changes across jurisdictions
Compliance advisor with actionable business guidance
Visual explainability map across 10+ APAC jurisdictions
Solo developer. Next.js + RAG architecture. Zero hallucination by design.
Win or lose, it's shipped. On to the next one.
Interesting direction. The part that feels most important in real workflows is not just whether the AI can generate something useful, but whether people can trust how it got there. In conversation-heavy work, context visibility and clear human handoff matter as much as raw model quality.
Exactly the framing I landed on. That's why every output in ReguLink ships with citation IDs + source URLs — not as a UI flourish, but because "trust the answer" only works when the user can click through to the actual statute text in seconds. Generation quality matters, but auditability is what makes it usable in real compliance workflows. Appreciate you naming the harder half of the problem.
Shipped solo mid-chaos. That's the whole thing right there. The scam interview was probably the best possible stress test -- if you can stay focused and still submit something worth looking at, you've got the mental muscle that actually matters for building long-term.
ReguLink sounds like a serious problem to be working on. Asia-Pacific regulatory fragmentation is a real pain for anyone trying to scale across that region.
Thanks Amanda — the scam interview was genuinely a useful filter. Either you flinch and lose half a day, or you close the tab and the deadline pressure suddenly feels manageable by comparison.
And yeah, APAC fragmentation is the whole reason this project exists. 10+ jurisdictions, laws shifting quarterly, most teams just hire local counsel per market and eat the cost. The diff engine is the part I'm most curious to see hold up in production — happy to share notes if you ever run into this space.