FindMeJobs Built for Real-Time Job Alerts
A developer built FindMeJobs using Cloudflare Workers and AI to scrape new job postings and send instant Telegram or email alerts the second they go live.
Stock photo for illustration only, not from the actual event
- Developer creates FindMeJobs to solve slow application times
- System scrapes new job postings every few minutes via Cloudflare Workers
- AI checks relevance before sending alerts via Telegram or email
- Architecture focuses on ultra-low latency for competitive job hunting
A common frustration in job hunting is discovering a great role only to realize it was posted three days ago and already has 400 applicants. By the time a vacancy appears in a standard feed, applicants are often simply too late to stand out.
To flip this dynamic, a software developer built a tool called FindMeJobs. Instead of humans constantly manually checking various job boards, the system monitors them continuously and messages the user the exact second a matching position goes live.
Users define their search criteria once, including keywords, boolean include or exclude rules, and locations. From there, an automated pipeline scrapes fresh postings every few minutes, filters them against the established rules, and runs the survivors through an AI relevance check.
Integrating artificial intelligence into the filtering pipeline addresses the common issue of noisy keyword search results. By evaluating semantic relevance rather than just matching words, the system ensures that professionals only spend time reviewing genuinely applicable opportunities.
Because being early is the core value proposition of the entire product, the architecture is engineered entirely around minimizing latency. Splitting the workload into distinct queue stages ensures that each component scales independently, meaning a slow scrape operation for one board never blocks notifications for another user.
- Scraping operates entirely on Cloudflare Workers and Queues on a cron schedule
- Queue segregation allows independent scaling of individual pipeline stages
- Isolated tasks prevent slow web scraping from delaying user alerts
The core application utilizes a standard T3-style setup, while the scraping and notification pipeline operates entirely outside the Next.js application framework. If a posting matches user criteria, an alert is dispatched via Telegram or email within minutes of the listing going live.
The product is currently live and operational, with an accompanying repository available for developers interested in examining the architecture and discussing the implementation details.
Source: Dev.to
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment