Job Hunting in the Age of AI
Run your job search like a data and AI project
Everything from the workshop, in one place. Score your skills against live market data, rebuild what's weak from one source of truth, and keep running the system long after class ends.
Taught by Jerome Hardaway · Founder, Vets Who Code · Google Developer Expert
Know the system. Then run it.
You don't have to finish it all this week. Start with Module 1 and work down. You can do this.
- 01
How the sausage is made
How AI-powered screeners read, score and pass you along.
- 02
How the data works
On every side: your skills, the market, the screener.
- 03
Your documentation, current
One source of truth that everything flows to and from: your portfolio.
- 04
Skills and resources
Skills, scripts and agents to run after today.
Measure everything.
If you don't track it, you can't fix it.
Never list what you can't defend.
Every line is a potential interview question.
Distribution beats talent.
The best candidate nobody sees loses to the average candidate who's everywhere.
Work the labs in order.
No labs run during the live session. Each module has one take-home lab and the tools to run it.
- M0
Intro, history and the lay of the land
Coding has been declared dead before. Every time, the level of abstraction went up. (No lab. Read the market history in the kit's references.)
- M1
Score your skills
Save 10 real postings for your lane, run step 1 of the chain on your resume, and mark each skill "can prove now" or "need proof."
- resume-scorer
- resume_score.py
- market_snapshot.py
- /score-resume
- M2
How the ATS parses, and how resumes are read
Run the parse test, decode one real posting (highlight, sort, count, map), then run a Screen Grade and note every qualification marked Not Met.
- resume-scorer
- screen_grade.py
- ats-screener
- M3
Your resume
Rebuild one lane resume from your portfolio record with the lift list, fill every [ask] and run the spec check. Save the PDF outside any repo.
- resume-rebuilder
- resume_rebuild.py
- resume_score.py
- M4
Your portfolio: the single source of truth
Run your first Friday Setup and curate five Brag Database entries. Pin only manicured repos, then run the GitHub score, the portfolio score and the sync check.
- friday_setup.py
- brag-curator
- github-scorer
- portfolio-scorer
- sync_check.py
- M5
LinkedIn
Score LinkedIn against the same target, fix it to match your portfolio, then run the two scores and the improvement plan.
- linkedin-scorer
- composite-scorer
- improvement-plan
- recruiter-screen
- M6
AI interviewers and defending every line
Run one mock AI screen on your rebuilt resume, grade every line Defended, Thin or Cut, rerun your two weakest answers, and build a Stack Defense Doc.
- /mock-screen
- ai-interviewer
- final-round-panel
- story-table.md
- stack-defense.md
- M7
Funnel, postmortems and offers
Start your funnel tracker, write one blameless postmortem, fill the offer matrix, and rerun the whole chain to see what moved.
- funnel.py
- postmortem.md
- offer-matrix.md
- career-scorecard
- /job-hunt
Seven steps. One live target.
Every surface is scored against the same live O*NET target. The career-scorecard agent runs all seven in order.
- Step 1
resume-scorer
Parse gate, lane score 0–100 and lift list; an A–D Screen Grade per posting.
- Step 2
linkedin-scorer
LinkedIn scored against the same target and your record.
- Step 3
github-scorer
Your profile and pinned repos, read from the keyless public API.
- Step 4
portfolio-scorer
Your site, the portfolio checklist and the sync check.
- Step 5
composite-scorer
Two reads, machine and human, plus a resume gate at 70.
- Step 6
improvement-plan
What to fix in each section, biggest gain first.
- Step 7
resume-rebuilder
Your resume rebuilt to spec from the portfolio, linted, defended, rescored.
Weights and the gate are a VWC design judgment, not research or any employer's. Change them with --machine-weights, --human-weights and --gate.
Get the kit.
A free, portfolio-first toolkit. Your portfolio is the base; your resume, LinkedIn and GitHub are scored against the market and rebuilt from it.
- Python 3.10+, standard library only. Nothing to pip install.
- GitHub CLI for Friday Setup. The GitHub scorer needs no token.
- Works in Claude Code, or load any SKILL.md into the AI tool you use.
Install as a Claude Code plugin
/plugin marketplace add Vets-Who-Code/ai-era-job-kit
/plugin install ai-era-job-kit@ai-era-job-kitOr copy into your user config
cp -r skills/* ~/.claude/skills/
cp agents/*.md ~/.claude/agents/
cp commands/*.md ~/.claude/commands/Then run
/job-huntAll public. All keyless.
Anyone can rerun every number. The scorers pull live on every run and stop if a source can't be reached. They never fall back to cached numbers.
O*NET database
Occupations and technology skills, with Hot Technology and In Demand flags.
O*NET OnLine
Look up the occupation code for your lane.
BLS Employment Projections
10-year growth, annual openings and median wage.
BLS Public Data API
OEWS wages, no registration needed.
Real postings
Your saved postings, or public Greenhouse, Lever and Ashby boards. What employers ask for now, including tools O*NET hasn't added yet.
Your data stays yours.
Keep resumes and your funnel tracker outside any git repository. Nothing in this kit sends, posts or applies for you. Every message is a draft you send yourself. Friday Setup redacts tokens, keys, emails and internal hostnames.
Start with Module 1.
Score your skills first. Everything else builds from the data. Free online, with limited in-person seats in downtown Minneapolis.