AI-native role, from discovery to deploy: coding agents with a human gate at every step; I decide architecture and contracts, and delivery goes through review and peer testing.
- Built the Seazone App, replacing 5 apps with a single login (IDP) because franchisees dropped property referrals without knowing which system held them, and the
@seazone-tech/uidesign system underneath it: semver, 718 tests across 30 components, accessibility measured with axe-core, and a 1-2 min CI per PR - Moved 15 phases of the contract termination workflow from manual pipelines to automation on the Django backend (Celery, ClickSign and Pipefy webhooks), with fail-closed financial verification: a debt found after signing becomes unrecoverable. Validated with the process owner on a golden set of 42 cases
- Built the Activity Center, a real-time chat over WebSocket (Django Channels) that became the single channel between franchises and the internal team and retired the third-party tool. Response time is now measured: median down from 2.7 to 1.6 business hours within a month, with 97% answered the same day
- Built solo, from PRD to deploy, the market intelligence platform covering 1,700+ cities and ~190K listings (Airbnb, Booking, Expedia, Decolar via AWS Athena), with a conversational agent over the data
- Own the governance of AI-assisted development across the team's repos (internal MCPs, agent rules, CI gates) and co-built the AI-driven workflow adopted as the standard across squads and SRE
- Operate what I ship: GitOps deploys (ArgoCD + Helm, HPA in production), debugging with Grafana (Prometheus, Loki, Tempo), backend tests parallelized (pytest-xdist), and the develop → staging → main flow locked in CI