Resume + LinkedIn — recruiter-ready

A recruiter spends 10-30 seconds on a resume on average. Optimise for that.

What works:
- One page (two if 8+ years experience)
- 3-5 bullets per role, each starting with a verb
- Each bullet has a NUMBER (improved X by Y%, shipped to N users, reduced cost by $Z)
- Top: name, link to GitHub + portfolio + LinkedIn
- No photo, no objective statement, no skills bar charts

The number per bullet is the single biggest upgrade. 'Built RAG pipeline' becomes 'Built RAG pipeline serving 5k queries/day, p99 latency 1.2s'.

Recruiters search LinkedIn by keyword. The headline + about are your SEO.

Headline (220 chars max):
- Role + tech stack + outcome focus
- 'AI/ML Engineer · LLMs · RAG · Python · shipping production agents'
- Not 'Aspiring AI Engineer'

About (3-4 short paragraphs):
- One sentence about who you are
- 2-3 sentences about projects + impact
- 1 sentence about what you're looking for
- Links to GitHub + portfolio

Update when your work changes. Don't let it drift.

PDF resumes from LaTeX or Markdown beat Word every time. Consistent typography, clean version control, easy diffs.

My default is rxresume / Awesome-CV / a small custom LaTeX template. Source on GitHub. Build one PDF per JD where the bullet order changes.

Don't rewrite your resume for every job. Recruiters can spot the over-tailored copy.

Do: rotate the order of your projects so the most-relevant one is first. Tweak the top bullet of each role to use the JD's verbs.

That 5-minute edit beats both 'send the same one' and 'rewrite everything'. Cover letters don't move the needle for ML roles. Time spent on a good portfolio repo will.

Day 87 done.

- Resume = highlight reel + numbers
- LinkedIn keywords matter
- LaTeX for clean PDFs
- Tailor lightly per JD

Tomorrow (Day 88): the ML interview prep loop that actually works.