problem

Job search discovery is noisy and fragmented. High-signal roles are buried across multiple sources.

approach

Harvester is a local job-search automation tool built to support a targeted remote job search. It started as a simple Python script that pulled public Greenhouse job listings into a CSV, filtered for remote roles, and preserved links, titles, companies, locations, and descriptions.

Over several iterations, it evolved into a configurable job-hunting pipeline.

capabilities

  • Separates noisy raw job discovery from a curated application tracker
  • Uses `boards.txt` as a configurable source list
  • Supports Greenhouse, Lever, Ashby, Workable, RemoteOK, We Work Remotely, Built In, Climatebase, Remote PM Jobs, Tech Jobs for Good, and direct Welcome to the Jungle/Otta job pages
  • Filters for Remote US roles
  • Filters for data product manager and technical program manager roles
  • Supports date filtering, defaulting to postings from the past 7 days
  • Preserves tracker state separately from scraped results
  • Allows manual promotion from `raw_jobs.csv` into `jobs_tracker.csv`

system

python * filtering * ranking * ai-assisted workflows * Codex

output

Terminal output from running the harvester pipeline
Spreadsheet output from harvested job results

next

Fit scoring * ATS tailoring * application tracking.

back to home