Portfolio Truthful Job-Search Pipeline

PYTHON · SQLITE · CLI · AGENTS · Personal project

job-hunt

Truthful Job-Search Pipeline

2026 2026 3 min read

Job searching is a data problem: many sources, inconsistent formats, and a CV that should be tailored every time. I built a pipeline that does the collecting and scoring, and a CV builder with a verifier that checks every title, date and number against a single source profile.

  • 1,350 roles collected from 71 companies
  • 154 hot of 1,245 scored roles flagged as strong matches
  • Verified CVs every claim checked against one source profile
job-hunt 2026 Mock · rendered from this write-up
  • Value Tailored CV per role
  • Time 1,245 Roles scored automatically
  • Effort Solo Personal tool
  • Quality 0 Invented claims
  1. Step 01 Official sources only
  2. Step 02 Score, then track
  3. Gate A CV builder that cannot lie
  4. Output Tailored CV per role
1,245 Roles scored automatically0 Invented claims1,350 roles collected from 71 companies154 hot of 1,245 scored roles flagged as strong matches

01 — Context

The problem.

Relevant roles are scattered across dozens of company career sites, each with its own format.

Tailoring a CV per role is slow, and AI tools happily “improve” a CV by inventing experience.

02 — Method

The approach.

The build, in the order it happened.

  1. Official sources only

    Greenhouse, Lever and Ashby public APIs, Workday career feeds and pasted links — no LinkedIn or Indeed scraping.

  2. Score, then track

    Each role is scored Hot/Warm/Cold against my profile and tracked in SQLite through to application status.

  3. A CV builder that cannot lie

    Tailored CVs are generated from one source profile, and a verifier checks every title, date, number and skill before output.

Flow · from the approach steps
  1. Sources PythonTyperSQLite
  2. Step 01 Official sources only
  3. Step 02 Score, then track
  4. Gate A CV builder that cannot lie
  5. Output Tailored CV per role

03 — Outcome

What changed.

  • 1,350 roles collected from 71 companies; 1,245 scored — 154 hot, 658 warm, 433 cold.
  • A tailored, verified CV and gaps note for each application.

04 — Screens

What it looks like.

Generated from the write-up rather than captured from a client's screen — the shape of the thing, not a leaked one.

Diagram: Greenhouse, Lever, Ashby and Workday feeds flow into a scorer and SQLite tracker, then a CV builder with a verification step outputs a tailored PDF and DOCX.

A personal pipeline that collects roles from official job-board APIs, scores them Hot/Warm/Cold, tracks them in SQLite, and builds a tailored CV per role — verified so it can never invent anything.

Ranked bars: 154 hot, 658 warm, 433 cold scored roles.

How the 1,245 scored roles split: 154 hot, 658 warm, 433 cold.

CV verifier stages: title, dates, numbers and skills each checked against the source profile.

The verifier: every title, date, number and skill is checked against the one source profile.

05 — Stack & role

Built with.

  • Python
  • Typer
  • SQLite
  • Greenhouse / Lever / Ashby APIs
  • python-docx
  • Headless Chrome
Role
Sole developer
Duration
2026
Client
Personal project
Period
2026

Got a report that takes two days to assemble?

That is usually a one-week fix. Tell me what you are reconciling by hand and I will tell you what I would automate first.

Start a conversation Back to the portfolio