Dubai, UAE  ·  🇵🇭 Philippines

💼  Team Lead – Web Development
👥  Leading & mentoring a team of 9+ developers
🏗️  Architecting scalable, real-time apps on Kubernetes, Docker & Cloudflare
🌐  Building multilingual, high-performance web platforms
🔐  Also own web admin: DNS, SSL, caching, security & fleet monitoring
🤖  Building AI-assisted engineering workflows — Claude Code, MCP servers & custom developer tooling
🎯  Obsessed with DevOps, infrastructure & production-grade systems
🏀  Sports enthusiast  ·  ✈️   Travel lover

• Real-time platforms — Kubernetes-hosted apps with MQTT-driven live data (scores, chat), Redis-backed caching/resilience, and MongoDB for high-traffic real-time state
• Real-time chat & community systems — Go microservices, gamified leveling/leaderboards, React clients, and Next.js moderation dashboards (bans, keyword filters, audit logs)
• Embeddable widgets — Next.js/React iframe-embeddable components (match grids, video players) with multi-brand theming and multi-language support (EN/VI/TH/ZH)
• Edge-cacheable microservices — Standalone NestJS APIs decoupled from WordPress, with Redis caching and observability dashboards
• Security tooling — In-house WordPress malware scanning, backlink monitoring, and fake-traffic detection deployed fleet-wide
• Database operations — MongoDB replica sets (deployment, TLS, keyfile auth, live migration and failover), MySQL, and Redis for cache and real-time state
• Fleet automation — Bash-scripted deployment, domain rotation, and cross-server database migration across the estate

Production problems I've owned end to end:

• Traced an API outage to the storage engine. Intermittent 503s and login failures came down to an aggregation fan-out saturating a MongoDB replica set's WiredTiger ticket pool — read queues hundreds deep. Reshaping the query cut steady-state load on the hot collection by roughly 40× and drained the queue to zero.
• Scaled a live message broker with zero downtime. Erlang misreads CPU topology on high-core hosts and crashes on start; after constraining it, I doubled the broker's core allocation live, mid-peak, at tens of thousands of connected clients — no restart, no dropped sessions.
• Migrated a production estate between servers. Databases, replica sets, brokers and the full site fleet moved hosts while staying online.
• Built the observability to prove all of it. In-house dashboards for uptime, edge traffic, replica-set health and storage-engine ticket pressure — because the outage that hurts is the one nobody is measuring.