Referral Playbook

How to Get a Referral for ML and AI Roles

Data science and ML roles pull hundreds of applicants, and almost every hire starts as a referral. Here is how to earn one from someone who can actually vouch for your code.

Find Your Referral Path →

Free · No sign-up · See results in ~10 seconds

Example - what you’ll see
in
Someone who works at your target company
🎓 Same university as you  ·  💼 Shared past employer
🔥 Strongest
in
A recruiter at your target company
🎓 Same university as you
🎓 Alumni

✍️ Ready-to-send intro“Hi - we both studied at [your school]. I’d love to hear about your path to a company you are targeting before I apply…”

… plus everyone else in your network who can put in a good word.

See who can refer you in - pick your target company:

Short answer: Find someone already on the target team who shares a school or past employer with you, then make the ask cost them 30 seconds. Send one project link that matches their stack, the exact role and req ID, and a two-line vouch they can paste into the referral form. Referrers say yes when helping you is easy and your work is visible, not when you send a long resume.

Why referrals decide who gets the ML interview

A single senior data science or ML opening can draw 300-plus applicants, and most resumes never get read by a human. A referral routes you into a separate pile that a recruiter is required to review, and it attaches a name to your file before anyone opens it.

For these roles the vouch matters more than in most fields, because managers want proof you can ship models, not just talk about them. A referrer who has seen your code or worked next to you gives the hiring manager a signal no cover letter can fake.

What to send the person who will refer you

Do not send a resume and a paragraph of hope. Package the ask so the referrer can act in one sitting and does not have to invent anything about you.

  • One 30-second artifact. Link a single repo, notebook, or deployed demo that matches their team's work, not your whole portfolio. Make the relevant thing the first thing they see.
  • The exact role and req ID. Internal referral forms ask for a specific opening, so paste the job link and the requisition number. Never make them hunt for which posting you mean.
  • Two lines they can copy. Write the vouch for them: what you built, the number it moved, and why this team. They will trim it, but a blank box is where referrals die.
  • Proof you fit the stack. Name the framework or model type they actually ship (PyTorch, Spark, ranking systems, LLM eval) and show one thing you have done with it. Generic 'data enthusiast' language reads as noise.

How to ask so it is easy to say yes

Lead with the thing you share, not the favor. Open with the shared lab, the class, the old team, or the mutual manager, then get to the ask in the next sentence. People vouch for people they can place.

Keep the request narrow and reversible. Ask about one role, offer to send everything they need, and make clear a no is fine. A scoped ask that costs 30 seconds gets answered; an open-ended 'can you help me break into ML' gets ignored.

Find the person who can actually vouch

A referral only counts when the referrer has standing on that team. An individual engineer or scientist on the actual team beats a recruiter or a distant VP, because their name on the form carries weight with the hiring manager who sits next to them.

The hard part is finding that person and confirming you share something real. FindWarmIntros does the finding-and-ranking for you: it surfaces the people at your target company who overlap with your school or past employers, ranks them by how warm the path is, and drafts the intro so you can lead with what you share instead of a cold ask.

Frequently Asked Questions

Can I ask for a referral from someone I have never met?
Yes, if you share something concrete like a school, a former employer, or a mutual manager. Lead with that shared thing, keep the ask to one specific role, and hand them a two-line blurb they can paste into the referral form.
Do I still need a referral if I have a strong GitHub?
Yes. A public project gets you taken seriously once a human reads your file, but a referral is what gets a human to read it. A strong repo plus a referral clears the resume screen far faster than either one alone.
Who is the best person to ask on the team?
An individual engineer or scientist on the actual team, not a recruiter and not a VP two levels up. They fill out the referral form directly, and their vouch sits closest to the hiring manager making the call.
What if they say the role needs more experience?
Ask them which one project would close the gap, build a scoped version of it, and follow up when it is done. That turns a soft no into a reason to reconnect, and a warm reintroduction later beats a cold application now.

Keep going

Find Your Referral Path →
Find your warm intro →