OpenAI Emerging Talent Program
The Emerging Talent track is built for people with 0 to 3 years of experience. You can apply as a sophomore or a recent grad. The application is free.
Tracks include applied research, product engineering, go-to-market strategy, and operations. You don't have to be a PhD candidate to get a callback. You have to show that you have already done something real.
Why this one matters OpenAI on a resume opens doors at every other AI company on this list. Even if you only get to round one, the application itself sharpens your story.
Direct Link
OpenAI Emerging Talent
openai.com/careers/emerging-talent/
Lead your application with the most measurable thing you've ever shipped. "Built a Discord bot that grew to 4,200 users" beats "Took CS 101 and got an A."
Scale AI — Technical Advisor & ML Research
Scale AI hires Technical Advisor and ML Research interns at around $50 per hour. Many roles are part-time and remote-friendly, so you can do this from your dorm without dropping a class.
The work is real. You are reviewing model outputs, evaluating reasoning quality, and contributing to the data pipelines that train Anthropic, Meta, and OpenAI's frontier models.
The angles
- Technical Advisor — domain experts (math, code, biology) reviewing model outputs
- ML Research Intern — full-time research positions on the in-house team
- SEAL / Outlier work — rolling, contract-based, fastest to start
Direct Link
Scale Careers
scale.com/careers
Scale's part-time advisor roles get filled as soon as they post. Set up a careers page alert and apply within 48 hours of any new listing going live.
NVIDIA University Recruiting
NVIDIA powers the entire AI revolution. Every model you've heard of trains on their chips. They are still rolling Summer 2026 internships across three tracks.
Where to focus
- Generative AI — research, model engineering, applied AI
- AI Computing — CUDA, inference, kernel work
- Developer Technology — tooling and SDK engineering for AI builders
Direct Link
NVIDIA University Recruiting
nvidia.com/en-us/about-nvidia/careers/university-recruiting/
Why NVIDIA hits different Every cycle of the AI economy runs through their hardware. An NVIDIA internship signals you understand the layer everyone else is building on top of.
If you have any CUDA, PyTorch optimization, or systems-level experience, lead with that. NVIDIA's university recruiters specifically search for low-level performance work.
Anthropic Fellows Program
This one is the sleeper pick. The Anthropic Fellows Program is a paid AI safety research fellowship with full mentorship and compute funding. They accept applications on a rolling basis with cohorts in May and July 2026.
You work on real safety research alongside the team that built Claude. Compute is provided. Mentorship is built in. You don't need a PhD.
Direct Link
Anthropic Fellows (Greenhouse)
job-boards.greenhouse.io/anthropic/jobs/5023394008
Backup Link
Anthropic Careers
anthropic.com/careers
For the Fellows application, the strongest signal is a writeup of an alignment, interpretability, or evals project you've actually run. A clean Colab notebook beats five lines on your resume.
The high-prestige backup list.
Apply to all five above first, then run through these. Same resume, two-line tweaks, send them all in one evening.
Adobe
AI/ML Engineer Intern
Generative AI work across Photoshop, Premiere, and Firefly.
adobe.com/careers
Salesforce
AI Research Intern (Palo Alto)
Foundation model research and applied science roles.
salesforce.com/company/careers
Microsoft
Microsoft Research / Applied AI
Filter "Student" + "AI" on the careers site. Dozens of open roles.
careers.microsoft.com
Meta
Research Scientist Intern
FAIR, Reality Labs, and ranking/recsys teams hiring across tracks.
metacareers.com
Perplexity
Research Residency / Intern
Rolling residencies. Smaller team, faster decisions.
perplexity.ai/careers
The high-ownership, high-pay route.
YC startups pay well, ship fast, and give interns real production code on day one. The YC Internships board lists hundreds of AI roles updated weekly.
Why this beats big tech for some students At a 10-person YC startup you ship something to real users in week one. At big tech you might still be onboarding by week three.
Live AI/ML internships on the board
- Dex — AI/ML Engineer
- Pulse — ML Engineer Intern
- Morph — ML Researcher PhD Intern
- Browser Use — SWE / ML Intern
- SafetyKit — AI engineering intern
- 14.ai — applied AI roles
- SimCare — ML internships in healthcare AI
Direct Link
YC Internships Board
ycombinator.com/internships
YC startups respond to direct outreach. Find the founder on X or LinkedIn, send a 3-line message with a working demo link. That hits 10x harder than the form.
Industry, research, and accessible programs.
These are the under-the-radar programs with the strongest yield-to-effort ratio. Most students miss them entirely.
Automotive AI
GM AI/ML Software Engineering Intern
Simulation, autonomy, and large-scale ML.
search.gm.com/careers
Enterprise
SAP iXp Business AI Research Intern
Search Indeed or sap.com/careers for Summer 2026 listings.
jobs.sap.com
Consumer ML
Pinterest ML Intern
Visual search, recommendation, and ranking models.
careers.pinterestcareers.com
Short-form
TikTok Applied Scientist / ML Intern
Global Frontier Tech Program runs multiple cohorts.
careers.tiktok.com
Gaming
Roblox ML Engineer Intern
Recommendation and ranking systems at planet scale.
careers.roblox.com
Research Lab
IAAIR Summer AI Internship
Applied AI and robotics — extended deadlines for 2026.
iaair.org
Quant / Hedge Fund
Balyasny Applied AI Summer Intern
Master's and PhD focused, with strong AI overlap.
bam.com/careers
Quant / Trading
Jane Street ML / Quant Research Intern
Top-tier ML research role with major AI overlap.
janestreet.com/join-jane-street/internships
How to apply with zero experience.
Every recruiter on this list scans your resume in 7 seconds. They are looking for one signal: can this person actually build something?
If you have no internships yet, you replace experience with proof. Three live projects beat any GPA.
The 7-day project sprint
1
Day 1 — Pick one problem
One. Not five. Something a real person you know would actually use. "AI tool that summarizes my class lectures" beats "general-purpose chatbot."
2
Day 2-5 — Build it with Claude Code or Cursor
Ship the ugly version first. You are not trying to impress designers. You are trying to prove you can take a problem to a working product.
3
Day 6 — Deploy it for free
Vercel, Streamlit, or HuggingFace Spaces. The link goes straight on your resume.
4
Day 7 — Write the loom
60-second loom of you using your own tool. Drop the link in your application. Most students never do this. It puts you in the top 5%.
The big unlock Recruiters don't expect you to have shipped GPT-5. They expect zero proof. The moment you provide any proof at all, you outrank everyone applying without it.
The 90-minute application sprint.
Block 90 minutes on your calendar. Do this in one sitting. Don't switch tabs to Twitter.
1
Open all five top links in tabs
Tesla, OpenAI, Scale, NVIDIA, Anthropic. Five tabs. No more. No less.
2
Update your resume to one page
Top section is projects with live links. Skills, experience, education in that order. Cut anything not AI/ML relevant.
3
Write one base cover letter
Three paragraphs. Why AI, what you've built, what you want to learn. Swap two lines per company.
4
Submit all five
Don't perfect. Submit. The version you send today beats the perfect version you never finish.
5
Follow the founder of one YC startup
Pick one company from the YC list. Find the founder on X. DM them with a working demo link. This is the lowest-effort, highest-signal move on the list.
"The version of you that applied beats the version of you that waited until your resume was perfect."