I put AI agents on real customers' WhatsApp lines, and I'm accountable when they get it wrong. Founding PM at Voltade in Singapore, engineer for five years before that. The work: evals, guardrails, failure taxonomies — what it takes to trust an agent with your business.
Conversational CRM for SMEs. WhatsApp-first inbox where a per-tenant agent triages, drafts, and replies; humans approve. 100+ active SME deployments, 230K+ AI interactions/day.
Self-serve, multi-tenant agent platform, in build now. Each tenant gets an agent-native WhatsApp helpdesk: inbox, event-driven agent harness, self-improving knowledge drive, automation. One pooled deploy; creating a tenant is an INSERT, not an orchestration.
App framework for AI coding agents. Bun + TypeScript + Drizzle with auth, Postgres, jobs, and an agent runtime baked in, so Claude Code gets working code on the first try. 100+ agents live across WhatsApp, Telegram, and Web.
Observability for AI agents. Designed a dashboard to monitor agent task runs, catch failures early, and surface execution quality issues before they compound.
Three posts that capture how I think about AI products: the system behind the headline numbers, the evaluation framework underneath it, and the guardrails that keep production agents on the rails.