8 Best Places to Find AI Talent in 2026

Job posts are one way to find an AI engineer. Public work is another, and you can read it before the first call. Engineers publish code and models in places you can search for free.

This list covers 8 places. For each one you get the price, the match time or the limits it publishes on its own site, read on 2 November 2026. For the vendors you can buy a placement from, see our guide to the best places to hire AI developers.

Disclosure: AY Automate develops AyHire and also ranks itself second in this list. Each entry lists its cons, ours included.

Best places to find AI talent: a brief overview

• AyHire: Best place to find AI talent assessed by shipped work. You describe the role in one sentence and it reads public GitHub work.
• AY Automate: Best when you want an engineer placed in your team instead of a search to run.
• GitHub: Best free place to read what an engineer built.
• Hugging Face: Best for engineers who publish models.
• Hacker News "Who is hiring?": Best free monthly thread for reaching engineers who read job posts.
• Toptal: Best for a vetted freelancer matched in about 48 hours.
• Upwork: Best for a scoped AI project with a price range you can see.
• LinkedIn: Best for engineers who keep their profile current there.

1. AyHire, best place to find AI talent assessed by shipped work

AyHire is developed by AY Automate. It assesses engineers by the work they shipped, not by their CVs, and Backbase and Sage use it.

You describe the role in one sentence. AyHire shows the searches it plans to run, reads public GitHub work (profiles, public repos, the READMEs of the five most-starred repos, commits, pull requests and activity), and scores each criterion Strong, Some, Unknown or Not shown, with a link to the work behind the verdict.

Key features

• A one-sentence brief becomes must-haves and a search plan you can edit before it runs
• A link to the repo, commit or pull request behind every verdict
• A call on each person: who to talk to first, one line of why, one thing to check
• Availability with its source, marked confirmed, likely or guess
• Works inside Claude Code, Codex and Cursor over MCP, and through an API
• Shortlists go to Greenhouse, Lever, Ashby, Workable and others, or to any tool through a webhook

Best for

• Founders and engineering managers hiring engineers who build in public
• Recruiters who must show why each person is on the shortlist
• Teams already working in Claude Code, Codex or Cursor

Pricing

• $99 a month for one person
• Team: $79 a month per person for 2 or more people, one bill
• 14-day free trial, no card needed. Cancel any time
• A free GitHub profile check at ayhire.com/engineer, no account needed

Pros

• Every claim links to the work it rests on
• Unknown counts for nothing either way, so nobody is marked down for private work
• You can paste your own list of GitHub handles or LinkedIn links

Cons

• It reads public work only. An engineer who works in private repos comes back Unknown and ranks lower
• It is weaker for roles with little public code
• It is software, not a placement. You run the interviews and make the hire

Try AyHire free for 14 days

2. AY Automate, best when you want an engineer placed instead of a search to run

If you would rather not source at all, AY Automate places an AI engineer from its own bench into your team. AY Automate works with 60+ clients. At Backbase it gave strategy advice and placed an engineer inside the team.

One client put it this way: "We needed a very specific role and AY Automate helped us get an AI engineer pretty fast." (Othmane Khadri, Founder, Earleads.com)

Pricing

• From $3K a month, cancel anytime
• 90-day replacement guarantee, no recruitment fee

Pros

• A proven AI system is set up in your company in hours, so work starts on day one
• Replacement terms are written down: 90 days, at no extra cost

Cons

• It is an ongoing monthly partnership, so it does not fit a small one-off task
• You meet the engineer we match, you do not choose from a pool of profiles

See engineer placement

3. GitHub, best free place to read what an engineer built

GitHub is where most public engineering work sits. Its people search matches what people write about themselves: a name, a bio, a location. Many strong engineers leave the bio empty, so searching for the work works better. Search repositories with qualifiers such as topic:, language:, stars:> and pushed:>, then read who wrote the commits (see AyHire's step-by-step guide to finding engineers on GitHub).

Pricing

• Free to search public repositories

Pros

• You see commits, pull requests and reviews before you contact anyone
• No tool fee

Cons

• Reading profiles by hand is slow, and the method runs out at volume
• A profile does not say whether the person is open to a new job

4. Hugging Face, best for engineers who publish models

Hugging Face lists 3,207,308 models on its models page as of 2 November 2026, next to datasets, Spaces and a Daily Papers feed. For roles in fine-tuning, inference or evaluation, the publisher of a good model is a candidate worth a look.

Pricing

• Free to browse

Pros

• Shows who built a model, dataset or demo
• Suits ML-leaning roles

Cons

• A model page names the publisher. It does not say whether that person wants a new job
• Weaker for engineers who build products rather than models

5. Hacker News "Who is hiring?", best free monthly thread for reaching engineers who read job posts

The whoishiring account posts "Ask HN: Who is hiring?" and "Ask HN: Who wants to be hired?" once a month. The November 2026 threads went up on 2 November 2026. Job seekers use the second thread to describe themselves.

Pricing

• Free

Pros

• Engineers in the "wants to be hired" thread have said they are looking
• No fee

Cons

• You post a comment and wait. Nothing screens who replies
• One thread a month, so you plan around the date

6. Toptal, best for a vetted freelancer matched in about 48 hours

Toptal's AI developer page says: "Typically, you can hire AI developers with Toptal in about 48 hours." It says fewer than 3% of the more than 200,000 people who apply each year are accepted. You can hire on an hourly, part-time or full-time basis.

Pricing

• No rate on that page. It says the cost "depends on various factors"
• A trial of up to two weeks. Toptal says you pay only if satisfied

Pros

• Toptal's talent matchers work with you on goals and needs
• Toptal states 98% of clients choose to hire after the trial, and a 4.9 out of 5 rating from 15,486 reviews. These are Toptal's own figures

Cons

• The AI page lists no rate
• You choose from the people Toptal puts forward

7. Upwork, best for a scoped AI project with a price range you can see

Upwork's AI engineer page says hiring one "generally costs $35-$60 per hour, depending on project scope and experience level." It also lists project ranges: $1,500-$5,000 for a chatbot or virtual assistant, $3,000-$10,000 for machine learning model development and $3,000-$12,000 for a generative AI application. The page shows a 4.8 out of 5 rating from 12,613 client reviews.

Pricing

• Rates set per freelancer, in the ranges above
• That page lists no fee percentage. Check Upwork's fee page before you budget

Pros

• Identity verification, payment protection and hourly tracking are listed on the page
• A price range for common AI projects before you post a job

Cons

• You read the profiles and judge the work yourself
• Quality varies by freelancer, so check past jobs

8. LinkedIn, best for engineers who keep their profile current there

LinkedIn is a professional network with a recruiting product, LinkedIn Recruiter. LinkedIn does not publish a price for LinkedIn Recruiter, according to AyHire's October 2026 write-up, which quotes LinkedIn saying pricing "varies based on your organization's needs".

Pricing

• Recruiter: no list price published

Pros

• Profiles list current employer and title

Cons

• A profile shows what a person says, not what they shipped
• Price only through a sales call

How to choose where to look

Do you want to search, or be sent candidates?

Search yourself with AyHire, GitHub, Hugging Face or Hacker News. Have candidates sent to you by Toptal or AY Automate.

Do you trust public work or a vetting stamp?

If you want to read the commits yourself, start with AyHire or GitHub. If you want someone else to have screened first, Toptal publishes its acceptance rate and AY Automate gives a 90-day replacement guarantee.

How do you want to pay?

Pay a monthly tool fee (AyHire, $99), by the hour (Upwork, Toptal) or through a monthly partnership (AY Automate, from $3K). If you want a side-by-side of three of them, see AyHire vs Toptal vs Upwork for AI engineers. For the process after you find someone, read how startups hire AI engineers.

FAQ

Where is the best place to find AI talent?

It depends on how you judge people. To read what an engineer shipped, AyHire is our pick: it reads public GitHub work and links every verdict to the repo, commit or pull request behind it. To have someone screen first, use Toptal or AY Automate.

Can I find AI engineers for free?

Yes. GitHub search, Hacker News threads and Hugging Face cost nothing to browse. AyHire has a 14-day free trial and a free GitHub profile check.

How fast can I find an AI engineer?

Toptal says about 48 hours to hire. The other places publish no match time, and a search through public work depends on how many profiles you read.

What do Toptal and Upwork cost for AI engineers?

Toptal publishes no rate on its AI page. Upwork's page says generally $35-$60 per hour.

Read the work before you book the call

Try AyHire free for 14 days. $99 a month. 14-day free trial, no card needed. Cancel any time.