Job Search Automation

Manual job applications are spray-and-pray - high volume, low signal, zero personalisation. A generic resume can't match every role without going over two pages or diluting the story. And finding the right person to contact at each company, writing a tailored email, customising the resume, and sending at the right time? Doing that manually for every application is a full-time job in itself.

I decided to treat my job search as a product problem. The insight was that I didn't need to apply to more jobs - I needed to apply better, with less manual effort per application. The first version was simple: paste a JD, generate a tailored resume and cover letter, apply. That alone saved time. But the real unlock was automating the sourcing layer so I could focus entirely on quality and personalisation at the output end, not on finding jobs manually. I spec'd the full pipeline before building, cut features that added complexity without improving outreach quality, and iterated on what actually affected reply rates.

A fully automated end-to-end outreach pipeline: Apify actors scrape jobs across 5+ portals (Hiring.cafe, Startup.jobs, Ashby, Greenhouse, RemoteOK), Apollo resolves decision-maker contacts with confidence scoring, OpenAI generates personalised emails with playbook enforcement, tectonic compiles per-JD LaTeX resumes, Gmail API creates drafts with OOO detection and scheduled sends, and Google Sheets tracks everything with A/B template testing across variants. Also built a Claude Code skill that runs the entire workflow automatically.

Fully automated daily outreach pipeline running in production. Every email is personalised, every resume is tailored to the JD, and sends are timed and tracked. The system itself is a live PM proof point: 0-to-1 build, iterated from a manual process to full automation, driven by a clear product insight about quality over volume.

Apify for multi-portal job scraping, Apollo API for contact resolution with confidence scoring, OpenAI GPT for email generation with template A/B testing, LaTeX + tectonic for per-JD resume compilation, Gmail API for draft creation and scheduled sends, Google Sheets as the tracking and analytics layer. Built a Claude Code skill (digiman) that orchestrates the full daily pipeline. OOO detection parses email headers and defers sends to the contact's return date.