Remote PM JobsRemote PM Jobs

How to Get an AI PM Job in 2026

What recent hires and the hiring managers who chose them are saying about the process, from the Skip podcast

Last updated

I'm a big fan of The Skip Podcast, in particular two recent episode, that, listened to together, form a complete picture of what AI product manager interviewing & hiring actually looks like in 2026. The episodes:

I highly recommend listening to these podcasts, as they give unique perspective from both the hiring and applicant perspective. But if you don't have the time, I've summarized the top takeaways for anyone looking to score an AI PM job in 2026:

  1. Apply through the website. It still works - all the hiring managers confirmed they still read applications through the website and frequently find candidates through the application. While connections matter, they only matter if that person has actually worked with you as a PM - not just a blind connection or friends / family.

  2. Build a targeted list before you're ready. Aim for 5–10 companies and do deep research on who you should connect with. By the time you're ready to apply, you'll have a good foundation.

  3. Make sure your cold outreach is worth opening. If you decide to reach out directly to the hiring manager, send along more than just a generic message. The best messages show depth of understanding of the product & initiative, like sending a prototype, a roadmap proposal, etc.

  4. In interviews, engage with the problem, not your past. Interviews are shifting to ~90% live working sessions. Show you can solve, not just narrate.

  5. Go deeper than the take-home asked. Do user research first. Defend every prototype decision. You don't have to know exactly how your prototype was built but you do have to understand the WHY behind every decision.

  6. Position your background as "I've seen what good looks like." Even if you don't have a high-value logo on your resume, as long as you can defend your product taste, you can position yourself well.

1. Apply through the website. It actually works.

Mackenzie at Netflix said most of her hires come from website applications. Sarah at Rippling: every application gets looked at. Sam at EvenUp: same. This is not a backup channel.

What's less talked about is when a referral is actually useful. Referrals carry weight only when the referrer has worked with you in a PM capacity — they can vouch for the work, not just the person. A referral from a friend, a family member, or someone whose only context is your LinkedIn profile can be, in Sarah's words, "rejected just as quickly" as a cold application. So if your strongest professional reference doesn't know your PM work, that referral isn't going to do anything for you.

The implication: stop assuming you need an "in." A clean website application from someone the hiring manager has never met can compete on the same footing as a friend-of-a-friend introduction.

2. Build a target list of companies before you're ready to apply.

Janie spent a week and a half having 50–60 conversations to narrow down to 5–10 companies. Ben described a 6-month, half-passive search, often three or four conversations deep with a company before he'd commit to interviewing. Julia tested the market while still at Pinterest, then took six months off to figure out what she actually wanted.

None of them ran a shotgun pipeline.

What that looks like concretely:

  • Source target companies via VCs you respect, recent fundraising news, and your existing network. The signal is where smart people are choosing to spend their time, not where the most jobs are posted.

  • Research each company deeply before reaching out. Identify the hiring manager, the recent product launches, the public statements from the founders. A weekend of research per company is the move. Knowing the company before you apply — let alone before you interview — is the single most important pre-application step.

  • If you know someone there, ask for a 15-minute call. If you don't, find a second-degree connection. If neither exists, move to cold outreach (next section).

A useful third channel worth flagging: exec recruiters. Julia found real value here, partly because they're not "on the company's team" the way an internal recruiter is. They want to place you — at this company or another one. Old recruiter messages sitting in your LinkedIn DMs from a year ago are warm intros. Reply to them.

If you can't identify target companies, you may need to take a step back and determine what you really want out of your career. Julia took a six-month break to ultimately determine she wanted to get back into IC work after managing at Pinterest. If you're not sure what you're aiming at, you may end up frustrated about your search or your next position.

3. When you cold-email a hiring manager, send something worth opening.

Janie gets 50+ cold outreach messages every week. The ones she opens have three things in common:

  • They're specific to her product. Generic "I'm interested in your company" gets skipped within a few seconds.

  • They include something the sender clearly built. A prototype, an observation grounded in evidence, a question that shows real research went in.

  • They make her learn something she didn't already know. If she finishes the email feeling smarter about her own product, she's writing back.

Sam at EvenUp describes the same pattern from the other side. When a candidate cold-emails him with "I think you should do X with this product, here's why," he wants to have that conversation immediately. That conversation is functionally interview round two or three, front-loaded by the candidate.

If your pedigree doesn't get you to the front of the line, a thoughtful product opinion on the company you want to work for absolutely can.

4. In interviews, engage with the problem. Don't lean on your past.

The most consistent shift across both episodes: interviews are moving away from "tell me about a time when…" and toward live problem-solving and working sessions. Julia's experience at OpenAI: roughly 90% of her loops were live working sessions on real problems the company was actively thinking about. The 10% behavioral content was almost a check that she'd been close enough to the work in her past role to operate as an IC.

Ben described the emotional pull most candidates feel before they figure this out:

"I'm talking to these companies that are AI native or looking for AI experts. I need to prove why I'm that AI person. Like, oh shoot, well, I haven't built an AI company from scratch. What stories am I going to tell? Should I explain that I have a patent because that shows that I'm an innovator?"

The conversations that actually moved his process forward were the opposite: not "let me tell you about my background," but "here's how I'd approach the problem you described — where is my thinking going wrong?"

What hiring managers are screening for in those moments is concrete. Sam at EvenUp said the bar has shifted from valuing the best prioritizer (someone who can navigate A vs B trade-offs cleanly) to valuing the frontier-pusher — someone who finds a way to ship both A and B faster, who looks for ways to have the cake and eat it too rather than negotiate scope down. Mackenzie at Netflix: depth of thought, ability to engage with assumptions, willingness to change a point of view when new information justifies it.

If you walk into an AI PM interview ready to defend your past, you'll lose to the candidate who's actively solving the problem in the room.

5. For the take-home, go deeper than they asked.

Take-homes are now standard. EvenUp and Rippling both use them in every PM loop. Both hiring managers spent significant airtime on what separates a strong take-home from a weak one — and the difference is almost never the prototype.

Sam at EvenUp on what stands out:

  • User research first, design second. Most candidates skip this. Even though deep-research mode on any modern AI tool can summarize Reddit threads, customer reviews, and public interview patterns in minutes. The candidates who come in with novel insights about the users — insights that weren't on the company's marketing page — stand out immediately.

  • Defend every decision under questioning. "I went down this path because…" If you get stuck explaining why a particular button exists or why the flow branches a certain way, your prototype was AI-generated structure you hadn't internalized. That's the most common failure mode he sees.

  • Go deep. If you don't have expertise in the domain, with AI tools, you can fast track your learning. Take the time to deeply understand the industry, the pain points, and the users.

Janie's take-home story is the model. Assigned a healthcare product she'd never seen, she spent the weekend reading the company's site, scouring YouTube for founder talks, building mocks, and writing a go-to-market plan.

Vibe-coding a prototype is table stakes now. Reading the prototype's outputs, catching its weird choices, defending the decisions — that's the actual work. It's product-managing the AI, not just using it.

6. Position your background as "I've seen what good looks like."

Pedigree is becoming less of a screen. Sarah at Rippling said pretty directly that she places little weight on brand-name companies on a resume. What she does care about: have you been close enough to high-quality product work to know what it looks like.

Two ways to play this depending on where you're coming from:

If you're at a big company that isn't AI-native: You probably haven't shipped AI features in production, but you have other signals worth leaning on. Promotions at a hard-judging culture travel — they signal you cleared a bar similar to the one AI companies are screening for. Lead with the specific work you owned, the decisions you made, the impact you can quantify. Don't apologize for the absence of AI experience; describe what you did see at scale.

If you're already in AI-native or AI-integration work: Lean directly into the taste calls you made. Don't say "shipped AI features" — say "owned the evaluation set for our agent product, made the call to stop chasing the long tail of model errors past a 95% quality bar, here's how I knew where to draw the line." Hiring managers screen for taste calls over AI experience, because the experience without the judgment is just incident response.

One counter-intuitive note Sam at EvenUp flagged: domain experience can be a negative. Candidates from the same vertical sometimes walk in assuming they know how the problem should be solved. That confidence in the wrong direction is harder to undo than no domain experience at all. If you're coming from inside the same industry as the company, lead with curiosity, not certainty.

What to do next

If you want to actually find the roles to apply to: