Start with the workflow
Before tools or models, understand the documents, calls, approvals, spreadsheets, and handoffs.
From law to enterprise salesto AI systems.
I'm Seth, a Singapore-based multidisciplinary operator-builder. I combine customer discovery, workflow mapping, technical scoping, and hands-on AI system building to turn messy operations into software people can actually trust.
"The interesting part of AI is not the demo. It is the operating layer around it: sources, workflows, review loops, and the human judgment that keeps the system useful."
GTMProductAI Workflows
LocationSingapore / San FranciscoFocusAI-native workStatusOpen to roles
How it started
I studied law at Cambridge and wrote my dissertation on machine learning in sentencing. That gave me the first lens I still use for AI work: the hard part is not just the model, but evidence, judgment, process, and the institutions around it.
Boxo and Airwallex taught me the enterprise side: selling technical products, mapping stakeholders, removing risk, and seeing how messy workflows become buying decisions. I learned to translate between operators, buyers, product teams, and technical constraints.
Salescraft was the hands-on turn: I built AI GTM systems for startups and learned how recurring edge cases become reusable workflow rules. That led me to San Francisco, where I worked with Sample Healthcare as a Forward Deployed Strategist on patient-intake and prior-authorization workflows. Seeing messy clinical documents, missing evidence, review queues, and approvals up close made the stakes of AI feel concrete. I carried that into Sunder, where I now build source-backed, reviewable systems across legal documents, CRM, company memory, and order intake.
How I Work
Stay close to operators
The best AI work comes from watching where people actually get stuck, not from guessing in a vacuum.
Keep review explicit
Citations, confidence, audit trails, and human approval matter when the output affects real work.
Translate across teams
I move between customer language, commercial reality, product shape, and technical constraints.
What I'm Looking For
I am looking for work where AI is close to real operations: messy inputs, human workflows, customer discovery, commercial outcomes, and systems that need to be trusted before they can be scaled. I am especially interested in San Francisco teams building AI Healthcare, applied operations, or GTM systems where product judgment and workflow depth matter.