Shipping OG from 0-to-1
In 2026 I'm the project leader for the launch of OG — a new prediction market platform spun out of Crypto.com as a standalone product. Concept to live in under two months. 200+ people across engineering, product, compliance, legal, risk, ops, and external partners — market makers, the works. CFTC oversight the entire way.
My job was to own delivery end-to-end — shaping the product alongside eng and design, driving the trade-off calls, and making sure 200 people building in parallel stayed aligned on the same version of the plan. A few things I took away:
1. Decide what "done" means before the first sprint
In regulated work, "feature complete" and "shippable" are different things. We split scope into three buckets early: launch-critical, fast-follow, and maybe-never. Everything got assigned an owner, and nothing moved buckets without a documented reason. When the timeline got tight — and it always does — that triage doc was what kept real-time scope trade-offs from turning into politics.
2. The critical path is usually people, not code
Engineering velocity was never the bottleneck. Legal review windows, compliance sign-off, market-maker onboarding, vendor integrations — those were. I stopped mapping the critical path by ticket and started mapping it by decision: who needs to say yes, by when, and what do they need to see to say it. That one reframing bought us weeks.
3. Write down every trade-off
I forced the trade-off calls through and made sure the whole team landed on the same version of the answer. If a decision got made and didn't get written down, it would be re-litigated three weeks later with less context and more tired people. So every call got a short summary: what we decided, what we deferred, and why. Boring, but it's what kept 200 people pulling in the same direction.
What compounds is the operating system underneath — clear scope, decision velocity, and a written record. That's what lets a 200-person effort move like a small one.