AI and Machine Learning Internships Summer 2027: Which Companies Are Hiring and How to Get In

TL;DR

• AI and machine learning internships for summer 2027 are open at dedicated AI labs (OpenAI, Anthropic, Google DeepMind, xAI), big tech AI divisions (Meta AI, Microsoft Research, NVIDIA), and AI-native startups. Most applications open between August and December 2026.

• Machine learning interns earn a median of roughly $88,400 annualized (~$42/hour) across 55 postings that disclose salary. Data science interns median lower at about $52,000 annualized (~$25/hour) across 54 postings. AI lab research interns at firms like OpenAI and Anthropic earn well above these medians.

• Python is the most in-demand skill, showing up in 82 of ~120 ML intern job descriptions sampled. PyTorch (46 mentions), C++ (29), TensorFlow (24), and SQL (16) round out the top five.

• You don't need a PhD or prior research experience. A solid portfolio with 2-3 projects, strong Python skills, and linear algebra fundamentals will get you interviews at most companies.

• This guide covers 30+ AI companies with direct links to our detailed internship guides for each.

New to the Extern world? Externships are short, remote, project-based programs that give you real work experience with leading companies. They're a great way to build your AI portfolio before applying to competitive internships.

So What Exactly Are AI and Machine Learning Internships?

What Does an AI/ML Intern Actually Do Day-to-Day?

AI and machine learning internships are 10-to-16-week programs where you work on real ML systems, from training models and building data pipelines to running experiments, co-authoring research papers, or shipping production AI features. That's what separates them from general software engineering internships. You're not building web apps. You're working on model development, data infrastructure, and applied intelligence.

The job titles vary, but most roles fall into a few buckets:

• Research intern: You run experiments, test hypotheses, and contribute to published papers. OpenAI, Anthropic, and Google DeepMind hire research interns who work on specific problems like alignment, reasoning, or multimodal models.

• ML engineer intern: You build and ship production ML systems. Think model serving, training pipelines, and performance optimization. NVIDIA and Meta hire heavily for these roles.

• Applied scientist intern: You use ML to improve product features. Amazon's Alexa team and Apple's Siri group bring on applied scientist interns to improve recommendation systems and NLP models.

• Data science intern: You analyze data, build predictive models, and surface business insights. This overlaps with ML work but leans more toward statistics and less toward deep learning.

If you want to see how AI roles fit into the broader tech landscape, check out our tech internships summer 2027 guide for the full picture.

AI Labs vs. Big Tech AI Teams vs. AI Startups: Where Do You Actually Fit?

Not all AI internships are the same. The experience you'll get depends on the type of company, and the differences are bigger than most people realize.

AI labs (OpenAI, Anthropic, Google DeepMind, xAI) are where the frontier research happens. Interns here often work on novel architectures, safety techniques, or capability evaluations. These roles lean academic. You'll read papers, design experiments, and potentially co-author publications. PhD students are common, but undergrads with strong portfolios get in too.

Big tech AI teams (Meta AI, Microsoft Research, NVIDIA, Google, Amazon) operate at production scale. Interns ship code that affects millions of users. The programs are more structured, with dedicated mentors, intern cohorts, and social events. If you want reliability and strong mentorship, this is the path.

AI startups (Together AI, Figure AI, Cerebras, Scale AI, Hugging Face, Runway) move fast. Interns wear multiple hats and often own entire features from day one. You'll learn quickly and build a lot, but the mentorship is less formalized. If you want speed and ownership, startups deliver.

The short version: research depth at AI labs, production scale at big tech, startup speed at startups. Pick the one that matches where you are right now.

Which Companies Are Actually Hiring AI/ML Interns for Summer 2027?

Here's where to look. We've grouped companies into three categories and linked to our detailed guides for each.

AI Research Labs

These are the highest-prestige AI intern roles. Competition is intense, but the experience is unmatched.

• OpenAI (Our Guide): Large language models, alignment, safety research

• Anthropic (Our Guide): AI safety, constitutional AI, frontier models

• Google DeepMind (Our Guide): Fundamental AI research, AGI-oriented projects

• xAI (Our Guide): Foundation models, reasoning

• Mistral AI (Our Guide): Open-source LLMs, efficient model architectures

• Cohere (Our Guide): Enterprise NLP, embeddings, retrieval

• Inflection AI (Our Guide): Conversational AI, personal assistants

Big Tech AI Divisions

These companies run dedicated ML/AI intern tracks separate from their general software engineering programs. Expect large cohorts, structured mentorship, and competitive pay.

• Meta (Our Guide): Open-source AI (Llama), FAIR research, computer vision

• Microsoft (Our Guide): Azure AI, Copilot, Microsoft Research

• NVIDIA (Our Guide): GPU computing, AI hardware, CUDA

• Google (Our Guide): Search AI, Cloud AI, TensorFlow

• Amazon (Our Guide): Alexa AI, AWS ML, robotics

• Apple (Our Guide): On-device ML, Siri, privacy-focused AI

• IBM (Our Guide): Enterprise AI, Watson, quantum ML research

• Intel (Our Guide): AI hardware, edge AI, model optimization

• Samsung (Our Guide): On-device AI, mobile ML

• Qualcomm (Our Guide): Mobile AI, edge inference, neural processing

AI-Native Startups and Applied AI Companies

Startups offer more hands-on responsibility, faster iteration, and often equity compensation on top of salary. These companies are building the infrastructure and applications that define the current AI wave.

• Scale AI (Our Guide): Data labeling, AI infrastructure

• Hugging Face (Our Guide): Open-source ML, model hub

• Databricks (Our Guide): Data + AI platform, MLOps

• Palantir (Our Guide): Applied AI, data analytics, defense

• Together AI (Our Guide): Open-source model training, inference infrastructure

• Figure AI (Our Guide): Humanoid robotics, embodied AI

• Harvey AI (Our Guide): Legal AI, LLM applications

• Tempus AI (Our Guide): Healthcare AI, precision medicine

• Cognition AI (Our Guide): AI software engineering, coding agents

• Perplexity (Our Guide): AI search, information retrieval

• Runway (Our Guide): Generative AI for video and image

• Character AI (Our Guide): Conversational AI, character models

• Cerebras (Our Guide): AI hardware, wafer-scale computing

• Waymo (Our Guide): Autonomous driving, perception AI

• Aurora (Our Guide): Self-driving technology, trucking AI

For a quick-reference view of all 32 companies with categories and focus areas, see the company table below.

When Do Applications Open for Summer 2027 AI/ML Internships?

The Real Recruiting Timeline, by Company Type

AI intern recruiting doesn't follow a single calendar, and honestly, that's what makes it tricky. Here's the general pattern:

August to November 2026: AI labs and large tech companies. Google DeepMind, Meta AI, NVIDIA, Microsoft Research, and Google typically open structured intern applications during this window. These programs have defined cohorts and close once filled. OpenAI and Anthropic tend to post roles on a rolling basis with shorter windows, so you'll want to check their career pages often.

October to January 2027: Big tech AI programs at full speed. Amazon, Apple, IBM, Intel, and Qualcomm ramp up AI intern hiring in fall. Many follow the same timeline as their broader SWE programs but route AI/ML candidates to specialized teams.

Rolling, year-round: AI startups. Companies like Together AI, Figure AI, Cognition AI, and Runway often hire interns when they need capacity rather than on a fixed cycle. Some only post openings two to four weeks before start dates.

The key takeaway: start looking now. If you're reading this in August 2026, you're right on time for AI lab and big tech applications.

Why AI Recruiting Feels So Different From Finance and Consulting

Here's the thing about AI intern recruiting that trips people up. There's no centralized deadline tracker. No "Super Day." Startups hire when they need people, not on a schedule. Even OpenAI doesn't announce application windows in advance.

So what do you do? Set up Google Alerts for career page updates, follow AI company accounts on LinkedIn, and check Greenhouse, Lever, and Ashby boards directly. Bookmark your top 10 companies and check every two weeks. It's not glamorous, but it works.

What Do AI and ML Internships Actually Pay?

Compensation by Role Type

AI internships are some of the highest-paying internships out there. Full stop.

Machine learning interns earn a median of roughly $88,400 annualized (about $42 per hour) across 55 postings that disclose salary, based on Extern Job Data Center data from August 2026.

Data science interns earn a median of about $52,000 annualized (roughly $25 per hour) across 54 postings that disclose salary.

AI lab research interns at companies like OpenAI and Anthropic reportedly earn well above these medians, with some research intern positions paying $10,000 to $15,000 per month according to Glassdoor and Levels.fyi reports.

One thing worth knowing: these figures are medians of published pay among postings that actually disclose compensation. Many companies don't list intern salaries publicly, so the real market range is broader than what these numbers show. Nobody publishes exact figures across the board.

And that gap between ML intern pay and data science intern pay? It reflects the specialized skill set. ML interns need deeper framework knowledge (PyTorch, TensorFlow) and stronger math foundations that data science roles don't require as heavily.

The skill bar below shows which skills appear most often in ML intern job descriptions, based on our analysis of approximately 120 postings.

The Programming Languages and Frameworks That Matter

Python is non-negotiable. It showed up in 82 of roughly 120 machine learning intern postings sampled in the Extern Job Data Center. That's the single most requested skill by a wide margin.

After Python, here's what counts:

• PyTorch (46 mentions): PyTorch has overtaken TensorFlow as the dominant ML framework. If you're going to learn one deep learning framework, learn PyTorch. It's the default at most AI labs and research groups.

• C++ (29 mentions): Critical for systems-level ML roles. Interested in NVIDIA, Cerebras, or any hardware-adjacent AI work? C++ is essential.

• TensorFlow (24 mentions): Still widely used in production environments, particularly at Google and companies that built their ML stack early.

• SQL (16 mentions): Important for data pipeline work and any role that involves working with large datasets.

• scikit-learn (11 mentions): The go-to library for classical ML. Good for feature engineering, model evaluation, and quick prototyping.

Also worth learning: JAX (Google DeepMind uses it extensively), NumPy, and pandas. Fluency in these signals that you actually work with data, not just tutorials.

The Math and Research Fundamentals That Won't Go Away

Frameworks change. Math doesn't.

The fundamentals that keep coming up in AI interviews and day-to-day work: linear algebra (matrix operations are the backbone of neural networks), probability and statistics (understanding distributions, Bayes' theorem, hypothesis testing), and calculus (backpropagation requires chain rule fluency, full stop).

For research roles, you'll also need the ability to read and implement papers, design experiments, and use LaTeX. A common AI lab interview question: "Can you implement a transformer from scratch?" You don't need every detail memorized, but you need to understand the architecture well enough to build it.

For production roles, software engineering fundamentals matter just as much. Version control (Git), testing, CI/CD, and deployment are all expected.

You don't need a PhD. But you need the math underneath the abstractions.

How Do You Break Into AI With Zero Experience?

Start With a Portfolio of 2-3 Real AI Projects

This is the single most important thing you can do. A portfolio of real projects beats a stack of certificates every time. And look, that's not some abstract advice. Recruiters at AI labs have said this publicly.

Here are three projects that show the skills AI companies actually hire for:

1. Fine-tune an open-source LLM. Take a model like Llama or Mistral and fine-tune it on a domain-specific dataset. This shows you understand model training, data preparation, and evaluation. Document your approach, hyperparameters, and results.

2. Build an AI agent. Create something that actually does useful work: a RAG chatbot that answers questions from a knowledge base, a code review assistant, or a data analysis agent. This is the hottest category in AI right now, and it proves you can build end-to-end systems. Check out our agentic AI project ideas for beginners for specific starter ideas.

3. Contribute to an open-source ML project. Pick a project you actually use (Hugging Face Transformers, LangChain, scikit-learn) and make a meaningful contribution. Even fixing bugs or improving documentation shows you can work in real codebases with other developers.

Host all projects on GitHub with clear READMEs, reproducible code, and results.

If you're new to building a technical profile, our guide to getting an internship with no experience walks through the full process step by step.

Kaggle, Research, and Other Ways to Build Credibility

Beyond portfolio projects, a few other signals can strengthen your application. But be strategic about which ones you invest time in.

Kaggle competitions. A top-10% finish in a relevant Kaggle competition shows applied ML skill in a way that's easy for recruiters to verify. Focus on competitions related to NLP, computer vision, or tabular data rather than niche competitions with small participant pools.

Undergraduate research. If you're still in school, approach a CS or ML professor with a specific project proposal (not a generic "I want to do research" email). Research experience carries real weight at AI labs.

AI hackathons. Events from companies like Anthropic, Google, or Meta give you concentrated project time and sometimes direct recruiting pipelines.

Selective online courses. The fast.ai course (free, practical, PyTorch-first), Stanford CS229 on YouTube, and Andrew Ng's courses are widely recognized. But don't over-credential. Two to three solid projects plus one competition result beats 10 certificates. Every time.

If you want structured AI experience before applying, Extern's remote Externships let you work on real AI projects with company mentors, building exactly the portfolio that AI recruiters want to see.

How Should You Prepare for AI/ML Technical Interviews?

Coding Rounds, ML Systems Design, and Research Presentations

AI technical interviews typically test three areas, and the weight of each depends on the company. Here's what to expect.

1. Coding interviews. Similar to standard software engineering interviews, but with ML twists. You might be asked to implement k-means clustering, write a training loop from scratch, or code a custom loss function. LeetCode-style preparation works, but supplement it with ML-specific coding problems.

2. ML systems design. This is the AI equivalent of system design interviews. You'll be asked to design a recommendation system, a fraud detection pipeline, a search ranking system, or a RAG architecture. The book Designing Machine Learning Systems by Chip Huyen is the best preparation resource out there. Think about data flows, model selection, evaluation metrics, and deployment trade-offs.

3. Research and behavioral interviews. You'll present a project, discuss a paper you've read, or explain your technical approach to a past problem. OpenAI and Anthropic lean heavily on this component. Big tech AI teams weight coding and ML design more equally.

Your prep plan: two months of LeetCode (arrays, trees, dynamic programming), one month of ML systems design, and at least two mock interviews where you present an ML project to a friend targeting similar roles. Is that a lot of prep? Yeah, it is. But these roles pay $42/hour for a reason.

Your Summer 2027 AI/ML Internship Game Plan

Here's your step-by-step checklist for landing an AI internship next summer:

1. Build 2-3 portfolio projects by fall 2026. Fine-tune a model, build an agent, contribute to open source. Get them on GitHub with clean documentation.

2. Learn PyTorch and review linear algebra fundamentals. If you only have time for one framework, make it PyTorch. Pair it with a refresher on matrix operations, eigenvalues, and gradient descent.

3. Set up job alerts on AI company career pages. Bookmark your top 10 companies. Check Greenhouse, Lever, and Ashby boards directly. Set Google Alerts.

4. Apply to AI labs August through November 2026 as postings appear. OpenAI, Anthropic, Google DeepMind, and xAI don't announce deadlines in advance. Apply as soon as you see a role.

5. Apply to big tech AI programs October through December 2026. Meta AI, NVIDIA, Microsoft Research, Google, and Amazon have more structured timelines. Get applications in early.

6. Prep for coding and ML design interviews. Start LeetCode now. Begin ML systems design prep in October. Schedule mock interviews for November.

7. Apply to AI startups on a rolling basis through spring 2027. Startups hire year-round. Keep applying even after you've submitted big tech applications.

8. Do a mock ML interview with a friend. Practice presenting a project, whiteboarding a system design, and coding under time pressure. One good mock interview is worth 10 hours of solo prep.

Ready to start building your AI portfolio now? Explore Extern's remote Externships for hands-on AI project experience with real companies.

What Comes After Your AI/ML Internship?

A strong AI internship opens doors that stay open for years. Here's what the post-internship landscape actually looks like.

Conversion rates vary by company type. Big tech AI teams convert 70-80% of strong-performing interns into full-time offers. AI labs are pickier, and a research internship doesn't guarantee a return offer, especially if the team is small or the research direction shifts. The numbers shift depending on who you ask.

Career paths from an AI internship include: ML engineer, research scientist, applied scientist, and AI product manager. Lateral moves into AI consulting, AI startup founding, and AI policy/safety are realistic too.

The market works in your favor. Experienced ML engineers are in extreme demand right now, which means strong internship performance often leads to multiple competing offers. That's not hype. That's what the hiring data shows.

For a broader view of where AI fits into the overall computer science job market and how tech internships work across specialties, check out our tech internships summer 2027 guide.

FAQs

Do I actually need a PhD to get an AI internship?

No. And that's one of the biggest misconceptions out there. While AI research labs like Google DeepMind and Anthropic do hire many PhD candidates, most also have undergraduate internship programs. OpenAI, Meta AI, NVIDIA, and many AI startups actively recruit undergrads for ML engineering and applied scientist intern roles. A strong portfolio of AI projects, solid Python skills, and math fundamentals (linear algebra, probability) matter more than your degree level for most positions.

How much do AI and ML interns actually get paid?

Machine learning interns earn a median of roughly $88,400 annualized (about $42 per hour) across 55 postings that disclose salary, based on Extern Job Data Center data from August 2026. Data science interns median lower at about $52,000 annualized. AI lab interns at firms like OpenAI and Anthropic reportedly earn well above these medians, with some research intern positions paying $10,000 to $15,000 per month according to Glassdoor and Levels.fyi reports. These are medians of published pay among postings that disclose, not market-wide medians.

What programming languages do I need for an AI internship?

Python is essential. It shows up in 82 of approximately 120 machine learning intern job descriptions sampled in the Extern Job Data Center. PyTorch has overtaken TensorFlow as the most requested ML framework (46 vs 24 mentions). C++ is important for systems-level ML roles at companies like NVIDIA and Cerebras. SQL rounds out the top five for data pipeline work.

What's the difference between an AI internship and a data science internship?

It comes down to what you're building. AI and ML internships focus on building, training, and deploying machine learning models, working with neural networks, LLMs, computer vision, or robotics. Data science internships focus on analysis, statistical modeling, and business insights using existing data. The skills overlap (both use Python and SQL), but ML interns need deeper math (linear algebra, calculus) and framework knowledge (PyTorch, TensorFlow). For more on the data side, see our data analytics internships summer 2027 guide.

When should I start applying for AI internships for summer 2027?

Now, if you haven't already. Major tech companies (Google, Meta, NVIDIA) typically open AI intern applications between August and October 2026. AI labs like OpenAI and Anthropic post on a rolling basis with short application windows. AI startups hire year-round. Unlike finance and consulting, there's no centralized recruiting calendar, so check career pages frequently and set up Google Alerts.

What AI projects should I build for my portfolio?

Focus on 2-3 projects that show you can build real AI systems, not just follow tutorials. Strong options: fine-tune an open-source LLM on a domain-specific dataset, build an AI agent (RAG chatbot, code assistant), or contribute to an open-source ML project (Hugging Face Transformers, LangChain). Host everything on GitHub with clear READMEs and reproducible code. A Kaggle competition result (top 10%) also helps, well, sort of. It's one signal among several, but recruiters do notice it.

Can I get an AI internship as a non-CS major?

Yes, especially in applied AI roles. Physics, math, statistics, and electrical engineering majors have the math foundations ML requires. Some AI companies also hire interns for AI policy, AI safety research, and technical writing roles that don't require engineering backgrounds. Your major title isn't the gatekeeper. What matters is whether you can show ML skills through projects and coursework.