Follow the JavaScript Mastery tutorial to learn how to build an AI agent that controls a real browser to automate multi-step web tasks
Add AI-powered job-match scoring to a recruiting tool by using GPT-4o to parse job descriptions and rank them against a profile
Generate a tailored resume that rewrites your base CV to highlight the skills most relevant to a specific job posting
Requires paid API keys for OpenAI, Browserbase, InsForge, PostHog, and Adzuna, plus a LinkedIn account configured for browser-session persistence.
JobPilot is a web application that helps technical job seekers find and apply to jobs automatically. Built as a tutorial project by the JavaScript Mastery YouTube channel, it connects to LinkedIn, Wellfound, and Y Combinator Jobs to search for positions, score how well each role matches your profile, and attempt to fill out application forms on your behalf using an AI-driven browser. The core idea is that you set up your professional profile once, and the tool does the browsing for you. It discovers job listings, parses each description with an AI model (OpenAI GPT-4o), assigns a match score, and stores everything in a filterable dashboard so you can review what was found. From that dashboard you can see the full job description, a breakdown of why the score is what it is, and controls to generate or tailor a resume for that specific role. The resume-generation feature creates a base resume from your profile data and then rewrites it to target a particular job description. There is also an experimental browser-automation path that tries to complete LinkedIn Easy Apply forms and external application-tracking-system forms on your behalf. The README is explicit that this part is experimental, and a separate report file in the repo covers where that automation currently works and where it does not. The project uses Next.js for the front end and server logic, TypeScript throughout, InsForge (a backend-as-a-service) for the database and user authentication, and Browserbase for running real browser sessions in the cloud. PostHog collects usage analytics. A YouTube tutorial walks through building the project step by step, and the code is structured around five named agent skills meant to show patterns for building AI-powered software. The README is aimed at developers learning agentic engineering, not at everyday job seekers looking for a finished product. Running it locally requires API keys for OpenAI, Browserbase, InsForge, PostHog, and Adzuna, plus a LinkedIn account configured for browser-session persistence.
← adrianhajdin on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.