GlobalChinaEV.
A multi-product platform tracking the global expansion of the Chinese EV industry — built from scratch, scaled to ~1M total views, and still growing.
The Chinese EV industry is one of the fastest-moving sectors in the global economy — yet English-language coverage was fragmented, shallow, and slow. Analysts, investors, and industry watchers lacked a single destination to track what BYD, NIO, Xpeng, and dozens of other players were actually doing across every market they entered.
The opportunity: build the definitive English-language hub for this space — with reliable data pipelines, quality editorial content at scale, and interactive tooling that no media outlet was offering.
AI Writing Agent
LLM-powered pipeline that monitors industry sources, synthesizes developments, and drafts publication-ready articles — enabling consistent editorial output without a full newsroom.
CMS Platform
Custom content management system built for speed of publication and SEO — structured to handle multiple content types (news, deep-dives, data stories) under a single editorial workflow.
Automated n8n Pipelines
Event-driven automation pipelines handling data ingestion, enrichment, content scheduling, and distribution — replacing manual workflows that would have required a dedicated ops team.
Interactive Charts Product
Live data visualisations of EV sales, market share, export volumes, and charging infrastructure — turning raw industry data into an interactive research tool for analysts and investors.
Airflow for orchestration, n8n for automation glue
Apache Airflow handles the heavy scheduled pipelines — data ingestion, batch processing, model inference runs. n8n fills the gaps with lightweight event-driven automations (webhooks, notifications, enrichment triggers). Keeping both let us avoid shoehorning either tool into a role it wasn't built for.
AWS-native infra for cost-effective scale
Content delivery on CloudFront, compute on ECS with Docker containers, storage on S3. The stack scales to traffic spikes without manual intervention and keeps infra costs proportional to actual usage — critical for a lean team.
LLM agent with editorial guardrails
The AI writing agent isn't fully autonomous — it produces structured drafts that go through a lightweight editorial review before publication. This keeps quality high while dramatically reducing the time per article, and avoids the hallucination risk of fully unsupervised generation.
Multi-product under one domain for SEO compounding
Rather than spinning up separate domains for charts, news, and data tools, everything lives under globalchinaev.com. Each product reinforces the domain authority of the others — organic traffic compounds faster than fragmented micro-sites ever could.
- —~1M total views across the platform — organic, without paid acquisition
- —Consistent editorial output at a fraction of the cost of a human newsroom
- —Live data products that competitors aren't offering in English
- —Fully automated ingestion-to-publish pipeline running 24/7 on minimal infra spend
- —Established as a reference source for English-language Chinese EV coverage