AI News Monitor
A real-time news monitoring dashboard for Indonesian oil and gas companies, replacing manual tracking with automated AI-powered article fetching and analysis.
Overview
News Monitor is an internal dashboard built during my time at SKK Migas Kalsul. The problem it solves is simple: monitoring news across multiple Indonesian oil and gas companies manually is slow, inconsistent, and easy to miss. This replaces that process with automated crawling, AI analysis, and a centralized view for stakeholders.
What I focused on
- RSS feed integration to continuously pull fresh articles from relevant media sources
- Google News integration
- A self-hosted Crawl4AI server to handle full article extraction beyond what RSS provides equipped with residential proxies
- AI-powered article summarization so stakeholders get structured insights without reading full pieces
- A Next.js dashboard that surfaces everything in a single view, prioritizing speed of reaction over depth of navigation
Outcome
The project replaced a manual monitoring workflow with something that runs continuously and surfaces analysis automatically. The most technically interesting part was getting the Indonesian-language sentiment model to produce reliable results on oil and gas industry text, which required careful prompt and pipeline tuning.
It was also my first time deploying a self-hosted AI inference service alongside a production web dashboard, which shaped how I think about separating ML workloads from application logic.