Handshake AI's screening doesn't look like the others. There's no single interview that decides everything. You apply once, wait to be matched, and then pass project-specific checks before you can start. Most people who stall do so at the first step, because their profile doesn't show what a project needs. Here's how it works and how to get through it.
What the Handshake AI screening actually is
Handshake's AI help centre describes the path in stages.
1. The application
You sign up through the Handshake AI Fellowship page. Along the way you:
- verify your phone number with a one-time code,
- answer work authorisation questions about the US, including whether you need visa sponsorship,
- fill in your expertise and background, and upload a PDF résumé,
- accept the programme agreements.
Handshake says the fellowship is open to students, graduates and professionals, and needs no prior AI experience. Most projects must be done from inside the US. Separate tracks exist for some PhD students elsewhere, so check eligibility before you start.
2. Matching
Handshake reviews applications on a rolling basis and matches people to projects by project needs, qualifications, experience and field. There's no fixed timeline. Submitting doesn't guarantee a project.
3. The invitation
If you match a project, you get an email describing the work, its length, the time commitment and the pay. You submit an interest form. Some projects also ask for a slide deck, portfolio or other work samples at this point. Then there's a short wait while candidates are reviewed.
4. Project assessments
This is the closest thing to an exam. Handshake says assessments are project-specific, so their format, length and pass mark vary. You may get:
- a knowledge quiz based on the project's training materials,
- a practice task or sample submission that the team reviews,
- a written or skills-based test in your field.
In most cases there's no retake. Failing one doesn't end your fellowship, though. Your profile stays active for other projects.
5. Identity and background checks
Before your first project you verify your identity through Persona, with a government photo ID and a selfie. Some projects also need a background check. For US fellows that runs through Yardstik and asks for details including your Social Security number. Fellows outside the US go through Checkr instead.
Handshake's help centre doesn't describe a live or AI video interview as a standard step. If a project adds one, the invitation will say so.
How to prepare
Profile checklist
- Use your full legal name, the same as on your ID. Handshake says nicknames and alternate spellings can get your account flagged.
- List your degree, field and institution exactly.
- Pick your areas of expertise carefully. They decide which invitations you see.
- Upload a clean PDF résumé that shows specific work in your field, not just job titles.
- Link only your own LinkedIn profile.
- Turn off any VPN when you use Handshake AI. Handshake lists it as a reason access gets paused.
Before an assessment
- Read all the onboarding material first. Handshake says the training materials are the best preparation, because assessments mirror the real work.
- Read every instruction twice. Then check your answer against it before you submit.
- Study the examples. Handshake says each project defines quality differently. The examples show you how.
- Ask early. If something's unclear, ask the project team before you start, not halfway through.
How to answer
- Show full reasoning. Say why an answer is right or wrong, step by step.
- Use your own field. If you're a chemistry PhD, write the problem a reviewer in your lab would respect. Concrete beats clever.
- Write it yourself. Handshake bans AI tools for work you submit, including grammar and style tools, unless it approves one in writing. It also asks you to keep your drafts and notes for 90 days in case it reviews your work.
- Don't reuse prompts from textbooks, Khan Academy, Chegg or other annotation platforms. Handshake doesn't count those as original.
Common reasons people fail
- A vague profile. No match means no invitation, and no invitation means no assessment.
- Skimming the training material and losing the one-shot assessment.
- Using an AI tool, even just for grammar. Under the policy, the work can go unpaid and you can be removed.
- Mismatched details between profile, Handshake account and ID.
- Location. Most projects require you to be in the US while you work.
What happens after
Once you pass onboarding, the project appears under Your Projects on your dashboard and you can start tasks. Handshake suggests starting soon, because availability changes quickly.
If you haven't been matched yet:
- Keep submitting interest forms in the Opportunities section of your dashboard. Handshake says that keeps you in matching.
- Finish any required assessments. Handshake calls a complete profile and finished assessments the most important things for getting invitations.
- You can verify your ID early, but Handshake says it won't raise your odds of being matched.
If you're accepted and still waiting, read passed the interview but no projects.
The catch
There's no clear moment where you pass. You can be accepted into the fellowship and never be invited to a project. Projects can start, pause or end at short notice, and new assessments can appear mid-project. The no-AI rule is also stricter than most platforms, grammar tools included.
More on the programme, including who can join, is on our Handshake AI page.
Some Apply links are referral links: we get a bonus if you're hired, it costs you nothing and the application is the same.