I solve marketplace growth problems at scale. 35M users, $180M revenue across 4 markets.
Leading product strategy at PropertyGuru, Southeast Asia's largest property marketplace. I write about product strategy, marketplace economics, and shipping AI features that solve real problems—not just demonstrate technology.

Background
I run product at PropertyGuru, Southeast Asia's largest property marketplace. I oversee the entire flywheel across four countries: the search and recommendation systems buyers use to find homes, and the monetization products that generate $180 million from property agents. My team ships features to 35 million monthly users while also building subscription models that keep 80% of paying customers renewing year after year.
The strategic tension is real. I raise prices 12-20% annually while maintaining retention above 80%. I optimize buyer discovery while building premium tools that help agents improve their visibility to those buyers. I balance four markets with different regulatory environments, trust expectations, and competitive dynamics. Most product leaders I meet own either experience or monetization, supply or demand. I'm accountable for the whole system, which changes how you see problems.
Before PropertyGuru, I led driver product at Ola during 2000% hypergrowth. We powered 1 million rides daily. I built driver features that the global ride-sharing industry adopted. Property and ride-sharing are opposite in almost every way: high-stakes vs. low-stakes, long consideration vs. instant decision, trust-dependent vs. convenience-driven. But the underlying marketplace mechanics repeat. I write about those patterns and what actually compounds when you're building at scale. If you're navigating similar tensions, this work might be useful.
What Makes Me Different
Licensed Pilot
Decision frameworks under uncertainty
Healthcare Researcher
Published cost-effectiveness analysis
Failed Entrepreneur
150k users, then learned what willingness-to-pay really means
Engineer + MBA
Fluency across functions
Writing
Getting the Most Out of Claude Cowork
30 minutes of setup turns Claude Cowork from generic assistant into a context-aware PM tool. Here are the 7 things I configured.
More Writing
Side Projects
These projects keep me close to the craft while my day job focuses on strategy and leadership. I build to test ideas and stay hands-on.
What I Learned
PMs don't want generic advice. They want specific guidance for their exact situation. This taught me how to design AI features that adapt to context instead of serving one-size-fits-all responses.
PMToolkit.ai
AI-powered product management learning platform
Project Details
Problem
Product managers lack integrated tools and structured learning resources for modern PM practices.
Solution
Built comprehensive platform with 20+ connected calculators, templates, and AI-powered insights for data-driven product decisions.
Impact
500+ active users, 10,000+ tool interactions, 4.8/5 user rating
What I Learned
The best productivity tools don't add features. They remove steps. I built this to eliminate the repetitive parts of my own workflow. Turns out, other PMs had the same problem.
ClipDrop
macOS CLI tool that saves clipboard content to files
Project Details
Problem
Product managers waste hours on repetitive copy-paste-save workflows when capturing research, feedback, and documentation.
Solution
Built macOS CLI tool with smart format detection, audio transcription via Apple Intelligence, and secret scanning to prevent credential leaks. Eliminates the tedious "Copy → Open app → Paste → Navigate → Name → Choose format → Save" workflow.
Impact
Open source, actively maintained with 14 releases
What I Learned
The biggest unlock for PM productivity isn't a better dashboard. It's meeting PMs inside the tools they already use. Building this as a Claude Code plugin instead of a standalone app meant zero adoption friction. It just works where you're already working.
PMCopilot
Claude Code plugin that gives product managers an AI copilot
Project Details
Problem
Product managers juggle dozens of disconnected tools for research, planning, and execution, losing hours to context-switching instead of making decisions.
Solution
Built a Claude Code plugin that turns the AI coding assistant into a full PM copilot. Ships with 4 custom MCP servers for PRD generation, competitive teardowns on real device simulators, sprint analytics from Jira, and stakeholder updates via Slack. Works in both Claude Code CLI and Claude Desktop with zero-config to full-stack setup tiers.
Impact
Open source, ships with 4 custom MCP servers, integrates with 10+ PM tools