Applied
Engineering
Results
samuel-torres.com
Role
Automating the repetitive parts of a job search — browsing listings and submitting applications — while keeping a local record of what was already processed. A personal project.
Core stack
TypeScript · Puppeteer · SQLite · Node.js · Chromium
Code
What it is
Built while job-hunting, to remove the repetitive, mechanical part of applying — opening a listing, filling in the same fields, submitting — while keeping a local record of what was already applied to, so the same listing never gets processed twice.
How it's built
- TypeScript running on Node.js, driving a real Chromium instance through Puppeteer rather than calling any undocumented API.
- SQLite as a local datastore tracking which listings have already been seen and processed.
- The bot can be stopped and restarted without losing track of progress — state lives in SQLite, not in memory.
Design notes
Puppeteer drives an actual Chromium instance and interacts with the page the way a person would — clicking, typing, waiting for elements — instead of reverse-engineering a private API, which is both more resilient to UI changes and closer to "automating what I'd do by hand."
Tracking processed listings in a small local database means the bot can be interrupted and resumed without re-applying to the same posting twice.
This was built for my own job search, run locally under my own account — not offered or distributed as a tool for others to run against LinkedIn.
Short version (portfolio card)
A personal job-application automation tool: TypeScript and Puppeteer driving a real Chromium browser, with SQLite tracking which listings were already processed.
Resume / LinkedIn bullets
- Built a browser-automation tool in TypeScript with Puppeteer to streamline a personal job search.
- Used SQLite to persist processed-listing state across runs, avoiding duplicate applications.
- Ran the automation against a real Chromium instance rather than an undocumented API, for resilience to UI changes.