Earn money by teaching AI models

Alignerr.com is a community of subject matter experts from several disciplines who align AI models by creating high-quality data in their field of expertise to build the future of Generative AI.

Alignerrs are individuals with unique skillsets who are building the future of Generative AI by creating high-quality data and evaluating performance of models.

During onboarding you’ll be equipped with the skills and knowledge you need to succeed in completing projects. You will then be added to our community where a leader will help you onboard and learn about how to successfully contribute to a range of projects.

Yes, we require contributors to provide their identification to ensure a safe platform and prevent bad actors. We maintain privacy standards in storing information. Please view our Terms of Use for more info.

Applicants should possess the required knowledge and specialized skills relevant to the areas they apply for. Our job postings provide detailed information on the specific requirements for each domain. Minimum qualifications typically include junior and senior undergraduates in the respective field. Preferred qualifications include graduate students, master's degree holders, and PhD recipients in the field. Proficiency in written English and effective communication skills are essential for all candidates. We are looking for individuals who are enthusiastic and dedicated to contributing to the development of cutting-edge Generative AI technologies.

We evaluate your resume and identify if you meet the minimum qualifications. We’ll then set up a screening call and administer an assessment to determine your suitability to become an Alignerr in the domain that you’ve applied for. Once your application has been approved, we’ll reach out to you to begin the process of onboarding.

The rates we offer are determined by your area of expertise and your level of qualifications as posted in the job description.

Yes, Alignerrs are independent contractors who can determine how many hours they want to work in a given week.

Alignerrs have varying schedules, with some work around 5-10 hours per week and others work up to 40 hours per week, depending on their availability. Your earnings are based on the number of hours you contribute each week, and there are no minimum or maximum hourly requirements. You have full control over your workload.

Alignerr’s projects aim to enhance AI models for improved safety, reliability, and accuracy. Your role involves assisting cutting-edge generative AI models in tasks such as coding improvement, data science synthesis, basic math and chemistry comprehension, and creative writing. Projects often involve addressing specific and diverse challenges while interacting with these models as they learn. The nature of your projects will vary based on your expertise, covering a range of topics. Tasks may involve generating training data to enhance models or evaluating model performance.

Contributors are freelancers, not employees, who can work when and where they want within the United States, Canada, Australia, and New Zealand. As an independent contractor of Alignerr, you’ll assume all liability for proper classification as an independent contractor or consultant based on applicable legal guidelines. Ability to freelance on Alignerr varies outside of the above geographies. Please refer to the job post in question for more details about accepted geographies.

No, as an Alignerr, you are a freelancer and you can work when you want, where you want, for however long you would like.

No, many contributors spend 20-25 hours per week on the platform, but as an independent contractor, you can choose how many hours you want to work a week.

This varies based on the task and type of project. We have a high number of Alignerrs available to perform generalist tasks, so this may result in longer wait times for relevant projects. Projects with high thresholds for domain expertise have rigorous requirements and are not as abundantly available, and when they are, often require a smaller team of Alignerrs. Project requirements are diverse and new projects come in frequently, when the right project for you comes about, you’ll be notified through email and Discourse.

We try to match the project as closely as possible with the skillset of each Alignerr. Prior labeling performance is the biggest factor in determining how Alignerrs get assigned. We measure both speed and quality, but the latter is the most important. Rate of data being accepted (as opposed to rejected or reworked) is key. Other skills that we take into consideration include:

• Domain specific assessment or test

• Verified education and/or work experience

• AI interview

• Human interview

• Prior labeling experience - both on and off of our platform

Alignerrs are independent contractors, not W2 employees. Work authorization is a process run by the federal government, and we cannot advise on the process.

Your ability to contract with Alignerr is dependent on the type of visa you have. Your visa must specify you are allowed to work and even then, you may still need authorization. If you are unclear on your qualifications, please contact your immigration advisor. Alignerr is unable to provide further guidance on your visa specifics.

Please note that we only accept US Driver’s Licenses, US State IDs, US Work Permits or US Permanent Residence Cards for identity verification. Identify verification is required for all contributors.

At this time, we do not sponsor work visas.

Yes, as an independent contractor, you can work when and for however long you like. Note that some projects do have specific educational requirements. Please reference the job posting in question if you have any questions about your eligibility based on your major or graduation date.

Alignerr is a product built by Labelbox for you to contribute aligning generative AI models and get paid. You will be using the annotation platform developed by Labelbox.

Yes! Alignerrs often improve existing or learn new skills. This is something to be proud of and share with the world!

Once you have signed up and onboarded, you will be added to Discourse. If you've passed your assessment and are having issues getting set up in Deel, contact support via email or Intercom (from within the Alignerr app). For any other issues, contact support using email or Intercom.

We assess all work activity prior to submission to clients and payment processing for Alignerrs. Suspicious or fraudulent activity may result in account suspension while we manually review the situation. This in turn may delay payments.

We will always send an email letting Alignerrs know specifically why they are being offboarded, but typical reasons include: 

• Misrepresentation of location (working one place, getting paid somewhere else)

• Misrepresentation of skills; cheating on assessments

• Cheating on labeling work (ie using GPT to write responses)

• Attempting to game the work timer (ie submitted 30 hours of work in a day)

• Consistently poor work output

• Unprofessional behavior

Fraud detection is a constantly evolving challenge. In order to mitigate fraudulent accounts, we implement the following methods for account verification:

• Rigorous onboarding process, including video interviews, background checks

• AI interviews

• Human interviews and experience verification

• Project level supervision

If your account was deactivated due to failing one of our account verification checks, you may experience delays in payment processing, as we need to manually verify all work completed to date. All completed and accepted work will be paid, however any fraudulent or incomplete work may not be paid. For updates on verification of work, you can contact support via email at support@alignerr.com

Project status will be posted in Discourse. We will update the users about the project, its performance, and status.

All projects go through a calibration phase where sample work is labeled and reviewed. Labelbox maintains performance metrics for both quality and speed to build a composite score for each task, project and all Alignerrs on each project. All work is reviewed for quality in a variety of ways, including:

• Consensus (comparison to fellow Alignerrs)

• AI critique

• Manual review (QA) for random samples 

Re-work rates are actively maintained for each Alignerr, project and client to assess quality and calibrate. A summary of typical performance levels include:

• Top performers: Quality data is consistently accepted in a timely manner. 

  • Opportunity for better pay rates and/or additional responsibilities like reviewer, project coordinator roles or Alignerr Connect

• Great performers: Quality data is accepted & paid

  • Work experience will lead to additional work project(s)

• Inconsistent / irregular performers: Quality data is accepted & paid

  • Higher rework rates and/or Inconsistent time to deliver may result in removal from project; limited future projects to ensure the quality of outputs for customers

• Fraudulent work -> rejected & not paid

If you have an unresolved issue, please contact support via email or Intercom (from within the Alignerr app). Escalations are manually processed which may take additional time and accurate payments will be issued as determined by the investigation. All Alignerrs are paid for completed work – even if it’s graded to be low quality. The only exception are the most extreme cases of fraud.