A full-stack WordPress platform built from scratch to aggregate, enrich, and publish Indian government job vacancies, exam results, and admit cards at scale. The system automates the entire content lifecycle — from fetching raw data via external APIs to enriching it with OpenAI’s ChatGPT, publishing posts, and instantly notifying subscribers via Firebase Cloud Messaging — with zero manual intervention.
Technologies & Frameworks
Core Implementations & Architecture
- Built a modular PHP content automation engine with a multi-stage pipeline: external API ingestion → post deduplication → ChatGPT enrichment (title, structured content) → WordPress publish
- Integrated OpenAI GPT API via Guzzle HTTP client to process raw job/exam JSON and generate SEO-optimized, structured post content programmatically
- Implemented WordPress cron scheduling and AJAX-driven manual sync triggers (full sync, delta sync, resync) with granular admin controls
- Built a Firebase Cloud Messaging (FCM) system with a custom MySQL token registry, per-device subscription management, notification delivery logs, click tracking, and auto-trigger on post publish via transition_post_status hook
- Developed a custom REST-style API fetcher with deduplication logic using source IDs to prevent content re-publishing across incremental sync runs
- Integrated Sentry SDK via Composer for production error monitoring with environment-aware DSN configuration
- Built utility modules: bulk CSV importer, IFSC bank code search (with JSON mapping), Indian Pincode lookup, and client-side image conversion tools
- Registered custom taxonomies and category-specific archive templates for latest-job, result, admit-card, and admission content types
Results
- Automated publishing of 400+ government job and exam posts with no manual content writing
- Eliminated manual content entry overhead entirely through the AI enrichment pipeline
- Push notifications delivered to subscribers automatically on every new post publish
- Centralized error visibility across production via Sentry with real-time exception capture
- Modular architecture allows independent scaling or replacement of any pipeline stage (fetcher, processor, notifier)