Loading page...Загружаем страницу...正在加载页面...
AtlasRepo

I tested AI Job Search for 1 week on a hypothetical job search — here's the honest take

Описание и применение

Тестирование AI Job Search показало, что инструмент автоматически сканирует вакансии, сопоставляет резюме и генерирует черновики сопроводительных писем, экономя 15‑20 минут на каждую заявку. Однако настройка занимает несколько часов, а скрипт LinkedIn нестабилен.

Задача

Job seekers often copy‑paste resumes into many forms, wasting time and risking privacy. A lightweight, local tool that automatically matches resumes to job postings and drafts cover letters would streamline this process.

Результат

The tool achieved 67% precision in surfacing fit jobs, produced useful cover‑letter drafts, and preserved resume privacy by running locally. Setup was time‑consuming and LinkedIn scraping was fragile, but the agent architecture proved highly reusable.

Рабочий процесс

  • 1. Prepare a structured CSV of resume details. 2. Configure API keys for OpenAI/Anthropic. 3. Run the CLI tool pointing to job sources (LinkedIn, Indeed, Lever). 4. Review the ranked output and generated cover letters. 5. Customize or refine the resume to improve missing keyword matches. 6. Submit applications using the drafted cover letters.

Доказательства

I spent 2 hours getting the dependencies right, configuring API keys for OpenAI/Anthropic, and getting the scrapers to play nicely with rate limits.

Источник

On the test set, AI Job Search surfaced 8 of the 12 "actually a fit" jobs in the top 15 results (67% precision).

Источник

The cover letters are not publishable, but they are good starting points.

Источник

  • Дата записи: 2026-09-15T05:01:17.725Z