Technical project manager at Lacak.io, backend roots. I ship operational systems for GPS tracking, fleet operations, and IoT — the kind of software where "it works" means trucks actually arrive on time. Most of my production work lives in private repos and client systems, so this profile is the short version.
MCP Lab — Postman for MCP servers. Connect to any Model Context Protocol server from the browser, browse its tools, run them from a schema-generated form, and read formatted results. Try it live.
fleet-mcp — a fleet telemetry MCP server: GPS positions, trips, geofences, and NMEA parsing for AI agents. Ships with a zero-setup simulated demo fleet, so you can point an agent at real-looking telemetry in one command. It's also the kind of server MCP Lab was built to test.
Shipped at Lacak.io, numbers from production:
- A fleet management system rollout with an estimated ~IDR 300M efficiency gain from a single-user baseline, with more upside as adoption grows
- Automation and BI pipelines behind SPX operational workflows — report processing 50% faster, expansion-monitoring deliverables 400% faster, order forecasting at >90% accuracy, and near-real-time SLA dashboards with >85% engagement
- ETL optimization (Python/SQL), FMS infrastructure tuning under real load, AI-based driver fatigue monitoring, a PLMS system for mining operations, and custom reporting for JNTC, JNTX, and BUMA
Python and TypeScript/Node day to day. PostgreSQL, MySQL, and Redis for data; Docker, Kubernetes, AWS, Prometheus, and Grafana for infra; React + Vite when something needs a front end. Currently going deeper on AI engineering — agents, MCP, and intelligent automation layered on top of the backend systems I already run.