AI Intern Radar — US & Canada

Less searching. More time for research.

Find AI/ML research internships and student research placements across the United States and Canada. Choose your degree, companies and research interests, then explore a clear visual shortlist with original employer links.

Chinese-language interface and reports are also supported.

Public Beta: Help us make the next version useful. Did you find a role worth applying for? What company or filter is missing? Share feedback in the Issues tab.

Explore the interactive demo — a dated sample, not a live job feed.

Get your shortlist in three steps

1. Start small. Keep the six default companies, US + Canada, PhD and 2027. Choose Master independently if it fits your search.
2. Run the Actor. Select Start. Runs use your own Apify account and platform allowance.
3. Explore your report. In Output, open Visual report. Refine filters, save interesting roles and visit the original postings. Filtering an existing report does not launch another scan.

Default sources: Cohere, Waabi, Together AI, Pangram Labs, Waymo and Databricks.

Built for focused internship searches

• Find your fit: filter by US or Canada, Master or PhD, company collection, research area, internship year and explicit workplace type.
• See the evidence: inspect original wording about locations, qualifications and timing. Unknown details remain unverified.
• Keep a personal shortlist: saved jobs and application progress stay in your browser. They are not synchronized across devices.
• Track observed changes: reuse a monitor_id and enable only_changes to identify changes since your previous live observation.
• Take the results with you: download Markdown, CSV or JSON from the visual report.

Fewer repeated requests

Complete live queries are cached for 24 hours by default, scoped to your account, monitor and exact query. Repeat runs can reuse those results without requesting the employer sites again.

A changed query, a new monitor, expiry, missing history or cache_max_age_hours: 0 triggers a fresh scan. Partial or failed scans are not reused. Cached reports preserve the original collection time and clearly state that sources were not checked again. They create no new change events or removal observations.

This is per-user query caching, not a shared daily job index. No collection schedule is enabled by default. Apify separately performs Store health tests.

Pricing

No additional developer fee. Users pay only for the Apify platform resources their runs consume, using their own available credit. Compute, storage and transfer still consume resources on a cache hit.

In one measured beta pair, the six-source live run and its cached repeat each displayed approximately $0.001 of platform usage. The cached run took longer in that pair, so we do not promise a guaranteed speedup or cost reduction. The verified benefit is avoiding repeated employer-site collection. Check your plan and allowance before running.

Outputs

Output	What it contains
Candidates or changes	Current matching records, or observed changes when requested. data_mode and data_collected_at distinguish fresh collection from cached or historical data.
Visual report (DASHBOARD)	A self-contained interactive HTML report.
English report / Chinese report (REPORT-EN, REPORT-ZH)	Markdown reports in both supported languages.
Summary (OUTPUT)	Counts, cache metadata, source errors and complete, partial or failed status.

The first observation establishes a baseline; it does not mean every role was just published. First-run changes-only output and cached changes-only output normally have no change rows. The visual report still contains the current shortlist.

Coverage and limits

The directory includes 100 selectable companies using public Greenhouse, Ashby and Lever boards plus four experimental scoped website adapters. This does not mean complete or equally reliable coverage.

Google, Microsoft, Amazon and NVIDIA use limited public searches or sitemaps. Errors and coverage limits are shown rather than hidden. Meta and Apple currently have official career links only. A company in the directory is not a claim that it is currently hiring interns.

No employer login, LLM, proxy or paid data API is required. Access restrictions are not bypassed. Degree labels reflect posting text, not verified eligibility. Remote work does not establish work authorization. Company tiers are editable preferences, not objective quality ratings.

Help improve the Beta

Use Issues to report missing companies, irrelevant or closed jobs, filter problems and feature requests. Include the company or posting URL, selected country and degree, expected versus actual result, and run ID when relevant.

Issues are public. Do not include API tokens, personal application details or private documents.

Information and liability disclaimer

AI Intern Radar provides public job information for informational purposes only. Information may be inaccurate, outdated or incomplete. We do not guarantee accuracy, completeness or timeliness and are not responsible for losses, omissions or application outcomes arising from use or reliance. Before applying, independently verify availability, eligibility, work authorization, deadlines and other requirements on the employer's official website.

Source code is MIT-licensed; employer text and third-party data retain their original rights. Follow source terms and apply through official employer links.